At the beginning of every year, I make predictions. At the end of every year, I score them. So here are a hundred more for 2016. Most of them are objectively decideable, but a few are subjective (eg “X goes well”) and are marked with asterisks.
1. US will not get involved in any new major war with death toll of > 100 US soldiers: 60%
2. North Korea’s government will survive the year without large civil war/revolt: 95%
3. Greece will not announce it’s leaving the Euro: 95%
4. No terrorist attack in the USA will kill > 100 people: 90%
5. …in any First World country: 80%
6. Assad will remain President of Syria: 60% [edit: called out as dumb, but I won’t cheat and change it]
7. Israel will not get in a large-scale war (ie >100 Israeli deaths) with any Arab state: 90%
8. No major intifada in Israel this year (ie > 250 Israeli deaths, but not in Cast Lead style war): 80%
9* No interesting progress with Gaza or peace negotiations in general this year: 90%
10. No Cast Lead style bombing/invasion of Gaza this year: 90%
11* Situation in Israel looks more worse than better: 70%
12. Syria’s civil war will not end this year: 70%
13. ISIS will control less territory than it does right now: 90%
14. ISIS will not continue to exist as a state entity [added: meant in Iraq/Syria]: 60% [edit: called out as dumb, but I won’t cheat and change it]
15. No major civil war in Middle Eastern country not currently experiencing a major civil war: 90%
16* Libya to remain a mess: 80%
17. Ukraine will neither break into all-out war or get neatly resolved: 80%
18. No country currently in Euro or EU announces plan to leave: 90%
19. No agreement reached on “two-speed EU”: 80%
20. Hillary Clinton will win the Democratic nomination: 95%
21. Donald Trump will win the Republican nomination: 60%
22* Conditional on Trump winning the Republican nomination, he impresses everyone how quickly he pivots towards wider acceptability: 70%
23. Conditional on Trump winning the Republican nomination, he’ll lose the general election: 80%
24. Conditional on Trump winning the Republican nomination, he’ll lose the general election worse than either McCain or Romney: 70%
25. Marco Rubio will not win the Republican nomination: 60% [edit: called out as dumb, but I won’t cheat and change it]
26. Bloomberg will not run for President: 80%
27. Hillary Clinton will win the Presidency: 60%
28. Republicans will keep the House: 95%
29. Republicans will keep the Senate: 70%
30. Bitcoin will end the year higher than $500: 80%
31. Oil will end the year lower than $40 a barrel: 60%
32. Dow Jones will not fall > 10% this year: 70%
33. Shanghai index will not fall > 10% this year: 60% [edit: called out as dumb, but I won’t cheat and change it]
34. No major revolt (greater than or equal to Tiananmen Square) against Chinese Communist Party: 95%
35. No major war in Asia (with >100 Chinese, Japanese, South Korean, and American deaths combined) over tiny stupid islands: 99%
36. No exchange of fire over tiny stupid islands: 90%
37. US GDP growth lower than in 2015: 60%
38. US unemployment to be lower at end of year than beginning: 50%
39. No announcement of genetically engineered human baby or credible plan for such: 90%
40* No major change in how the media treats social justice issues from 2015: 70%
41* European far right makes modest but not spectacular gains: 80%
42* Mainstream European position at year’s end is taking migrants was bad idea: 60%
43. [Duplicate removed]
44* So-called “Ferguson effect” continues and becomes harder to deny: 70%
45. SpaceX successfully launches a reused rocket: 50%
46* Nobody important changes their mind much about the EMDrive based on any information found in 2016: 80%
47. California’s drought not officially declared over: 50%
48. No major earthquake (>100 deaths) in US: 99%
49. No major earthquake (>10000 deaths) in the world: 60%
50. Occupation of Oregon ranger station ends: 99%
1. SSC will remain active: 95%
2. SSC will get fewer hits than in 2015: 60%
3. At least one SSC post > 100,000 hits: 50%
4. UNSONG will get fewer hits than SSC in 2016: 90%
5. > 10 new permabans from SSC this year: 70%
5. UNSONG will get > 1,000,000 hits: 50%
6. UNSONG will not miss any updates: 50%
7. UNSONG will have higher Google Trends volume than HPMOR at the end of this year: 60%
8. UNSONG Reddit will not have higher average user activity than HPMOR Reddit at the end of this year: 60%
9. Shireroth will remain active: 70%
10. I will be involved in at least one published/accepted-to-publish research paper by the end of 2016: 50%
11. I won’t stop using Twitter, Tumblr, or Facebook: 95%
12. > 10,000 Twitter followers by end of this year: 50%
13. I will not break up with any of my current girlfriends: 70%
14. I will not get any new girlfriends: 50%
15. I will attend at least one Solstice next year: 90%
16. …at least two Solstices: 70%
17. I will finish a long blog post review of stereotype threat this year: 60%
18* Conditional on finishing it, it won’t significantly change my position: 90%
19. I will finish a long FAQ this year: 60%
20. I will not have a post-residency job all lined up by the end of this year: 80%
21. I will have finished all the relevant parts of my California medical license application by the end of this year: 70%
22. I will no longer be living in my current house at the end of this year: 70%
23. I will still be at my current job: 95%
24. I will still not have gotten my elective surgery: 80%
25. I will not have been hospitalized (excluding ER) for any other reason: 95%
26. I will not have taken any international vacations with my family: 70%
27. I will not be taking any nootropic daily or near-daily during any 2-month period this year: 90%
28. I will complete an LW/SSC survey: 80%
29. I will complete a new nootropics survey: 80%
30. I will score 95th percentile or above in next year’s PRITE: 50%
31. I will not be Chief Resident next year: 60%
32. I will not have any inpatient rotations: 50%
33. I will continue doing outpatient at the current clinic: 90%
34* I will not have major car problems: 60%
35* I won’t publicly and drastically change highest-level political/religious/philosophical positions (eg become a Muslim or Republican): 90%
36. I will not vote in the 2016 primary: 70%
37. I will vote in the 2016 general election: 60%
38. Conditional on me voting and Hillary being on the ballot, I will vote for Hillary: 90%
39* I will not significantly change my mind about psychodynamic or cognitive-behavioral therapy: 80%
40. I will not attend the APA meeting this year: 80%
41. I will not do any illegal drugs (besides gray-area nootropics) this year: 90%
42. I will not get drunk this year: 80%
43* Less Wrong will neither have shut down entirely nor undergone any successful renaissance/pivot by the end of this year: 60%
44. No co-bloggers (with more than 5 posts) on SSC by the end of this year: 80%
45. I get at least one article published on a major site like Huffington Post or Vox or New Statesman or something: 50%
46. I still plan to move to California when I’m done with residency: 90%
47. I don’t manage to make it to my friend’s wedding in Ireland: 60%
48. I don’t attend any weddings this year: 50%
49. I decide to buy the car I am currently leasing: 60%
50. Except for the money I spend buying the car, I make my savings goal before July 2016: 90%
Other people doing yearly predictions with probability: Against Jebel al-Lawz, Anatoly Karlin, Old Lamps, Garrett Peterson. If you’re doing this and I missed it, let me know and I’ll add you in.
This one is missing a number?
Howso? Asterisk means it’s subjective.
It’s missing a probability; the other subjective ones aren’t?
Oh! That kind of number!
I added it in.
That’s confusing. Maybe I’m not understanding what the Ferguson Effect is, but surely crime rates can be measured. Is it whether a protest counts as major enough to cause it subjective?
Crime rates aren’t very meaningful in this case. The problem is that 2014 had the lowest violent crime rate ever for the US. So any increases after 2014 are hard to pin down on ‘things being bad’ vs ‘things not being unrealistically good’.
But you could still compare the increase in areas where protests occurred to those where it didn’t, or areas where the police department’s policies have changed in response to the protests vs those where they didn’t. Or is the idea that the protests cause increased crime rates everywhere because all police are worried about it?
Broggly, http://www.unz.com/isteve/nyt-editorial-false-alarms-about-a-national-crime-wave/?highlight=homicide may be of interest.
@Broggly: Just where the protests happened probably isn’t meaningful since its national coverage, but looking just at places where policy changed in 2014/2015 might work if we could get the data.
You seem to be more optimistic on Hillary and pessimistic on Trump compared to the Scott A Drawn With A Very Fine Camel Hair Brush, I wonder why.
The other Scott A believes there’s a 99% of a Trump landslide. He’s either deluded or else a genius way beyond all of our comprehension.
More specifically, he’s working from a model where good persuasive abilities might as well be magic and someone who has those abilities can do things that the rest of us would consider impossible. I don’t share his model – I think persuasive abilities are helpful but won’t necessarily overcome a bunch of other unfavorables. As a result, I think Trump – who I acknowledge is a competent persuader – will still be sunk by the fact that most people genuinely don’t like him and associate him with toxic ideas like racism and populism, plus he has little governing experience. If the other Scott A is right, I will happily update my model, having learned something really interesting.
As for Hillary, basically I just can’t see who else besides her would win. Sanders’ chances are being exaggerated; he could win Iowa and NH because they’re both very white states, but as I discussed before, he has minimal support among minorities and minorities are too big a share of Democrats in most states for that to be at all sustainable. The Republican field is a mess; Rubio could do okay in a general but doesn’t seem to have any nomination moment, and I thought Bush could do okay in a general but even his own party doesn’t seem to like him.
Populism is generally seen as a toxic idea in your model? I’d be interested to hear more about that. It’s a mostly-neutral-if-vaguely-disreputable one in mine.
Ignoring whether it’s actually toxic, it’s seen as toxic by a lot of voters and especially by opinion-setting elites.
Surely part of the lesson of Trump’s success so far is that opinion setting elites have less influence than most of us expected. And it isn’t just with Republicans–the Democrats also have a candidate doing much better than he should be.
While that’s of course true, primaries (actually _before_ primaries) are too early to write off the historically significant status of opinion setting elites entirely (to say nothing of “lots of voters”). It’s entirely plausible that general election voters (who are more centrist than primary voters, as a group) would exhibit a lower affinity for populism as well as being influenced more by elites, since as a group they’ve made up their minds to a far lesser degree than primary voters.
The opinion-setting elites could be responsible for Trump’s appeal. When a sizeable wedge of potential voters are constantly being lectured about how horrible they are merely for breathing, they’re quite likely to decide they may as well be hung for a sheep as a lamb and support The Big Bad Bogeyman.
Whether that translates into actual votes when the election really counts, I have no idea.
I think it’s more the xenophobia that makes people hate Trump than populism. Bernie Sanders is very populist (and not even super pro-immigration) and he doesn’t draw the same kind of hate.
How can populism be toxic in a general election? Isn’t populism, by definition, that which appeals to popular sentiments?
> When a sizeable wedge of potential voters are constantly being lectured about how horrible they are merely for breathing…
Which are you referring to?
Somebody define “populism,” please.
“Trying to appeal to the people who oppose me in relation to matters in which my view is in the minority.”
Wikipedia: Nonetheless, in recent years academic scholars have produced definitions of populism which enable populist identification and comparison. Daniele Albertazzi and Duncan McDonnell define populism as an ideology that “pits a virtuous and homogeneous people against a set of elites and dangerous ‘others’ who are together depicted as depriving (or attempting to deprive) the sovereign people of their rights, values, prosperity, identity, and voice”.
If you asked me to classify a long list of policies by level of populism, then tried to work out what definition fits the way I actually use the term, I think you’d end up with something like “taking positions and advocating policies that are not reasonable based on the available evidence, but are emotionally appealing, and defending those positions and policies with appeals to emotion rather than with logic and evidence”.
Another way of expressing this is that if a domain expert says “well, he’s wrong, but explaining why would take me a while”, the politican is not (currently) being a populist; whereas if a domain expert gets angry about dodgy statistics, fallacious arguments and outright lies, the politician is being populist.
“taking positions and advocating policies that are not reasonable based on the available evidence, but are emotionally appealing, and defending those positions and policies with appeals to emotion rather than with logic and evidence”.
I think it’s the defense rather than the positions that makes it populism. There is no reason why some of the positions taken by a populist can’t be ones that are reasonable based on the available evidence.
“whereas if a domain expert gets angry about dodgy statistics, fallacious arguments and outright lies, the politician is being populist.”
The fact that a politician is dishonest doesn’t make him a populist—he could be dishonest for any of a variety of different reasons.
I think of it as something like “politicians who give the distinct impression they couldn’t sketch out a model of how they think the economic/political/social system works on a napkin if you held a gun to their head.”
I don’t agree with Marxists, but they do have some general concepts in there about class conflicts and bargaining power and etc. Populists don’t. They tend to propose ad hoc laws to fit specific policy problem and if you point out a negative consequence of that law, they’ll suggest another law to just fix that consequence.
Obviously all successful political parties have some incoherent policies, this is a matter of degree.
@ Tracy W
>> [Someone Tracy quoted thinks] of it as something like “politicians who give the distinct impression they couldn’t sketch out a model of how they think the economic/political/social system works on a napkin if you held a gun to their head.”
> [Tracy W said] I don’t agree with Marxists, but they do have some general concepts in there about class conflicts and bargaining power and etc.
Tracy, what are some other factions whose proponents can sketch this on a napkin (and whose politicians would, absent the gun)? Libertarians? Monarchists?
> Populists don’t. They tend to propose ad hoc laws to fit specific policy problem and if you point out a negative consequence of that law, they’ll suggest another law to just fix that consequence.
I suspect that this is a good thing. Sort of, yanno, evolutionary. Few laws get passed without a lot of support and little opposition; and if and when a bug turns up, there’s not much popular opposition to de-bugging the law (or letting it wither away).
@ Chrysophylax, David Friedman
Those sound like the sort of definitions that would be composed by elites, and news to me. I thought ‘populist’ was about positions, policies, laws that might get made, and practical consequences for (excuse the expression) “the little guy”. The tenants vs the landlords. The debtors vs the lenders. Etc.
If ‘populist/ism’ were about appeals to emotion, then it would apply to any kind of emotion, including ‘Don’t be like those stupid hillbillies who want lower rents and lower interest rates; be like Us Ivy Leaguers.’ I don’t think that’s what the quoted definitions mean, though.
I’ve always thought of populism as being specifically responding with short term, simple mechanism of effect action to the political issues du jour without analysis of actual effectiveness. Announcing action to solve what people want solving the way they think they want it solved, without filtering on cost, side effects, basic viability, or whether the solutions actually work.
Examples being things like:
1) Responding to a specific crime by passing a new law to Make That Crime Not Happen (sometimes named after the person) without much consideration of rates or whether the proposed law makes sense or would actually do anything about the crime.
2) Responding to bad people doing bad things by announcing that you’re going to get rid of them, without concern for whether they’re actually violating any laws.
3) Responding to shifts in public opinion of programmes by starting/stopping them wildly (and expensively).
4) Responses to moral panics du jour in the form of illegalising things wildly, without reference to actual rates of occurrence, often with poorly written ambiguous laws.
5) Responding to unemployment by declaring that you’re going to hire everyone/pay employers to hire everyone.
6) Responding to difficulties people have competing in the market through simply taxing or barring imports, or direct price controls.
It’s an insult because it basically means “You’re not thinking things through”. It’s not automatically maximally the case in a democracy, because voters tend to vote on outcomes- how well the economy is doing, etc- rather than on adherence to their specific demands, so an intelligent leader can find it a lot better to ‘manage’ stupid demands and focus on good outcomes at least once they’re in office. It’s definitely true that things you don’t like will look more populist than things which you don’t, but it’s also a recognisable type of behaviour.
Whether Trump *is* particularly populist I’m not sure of. “Build a wall” and the protectionist ideas definitely seem to qualify, and I’d throw the Muslim ban in too since I understand it to be infeasible-and-he-should-have-known-that. But while he has a few really strong examples the others have bunches of mildly ones and it’s nothing approaching the archetypes set by the more dysfunctional countries out there.
For extra lulz, try to define the difference between populism and democracy or a democratic spirit.
The funniest example is the poor hating on the rich, it can be a democratic thing when leftie elites approve of it, usually when it has no ethnic undertones and maybe has undertones of universal utopia, and populism when someone with a rightie smell spreads it, often with ethnic undertones and has more tribal undertones we get their stuff, we will be happy, everybody else can bugger off.
Somebody define “populism,” please.
Well, the Populist Party Platform of 1896 is at
Dunno how many later parties that chose to use the term ‘Populist’ are closely aligned with the 1896 party or with each other, but here is some indication of what some of them are/were about.
American Populist Party
Individualized political contributions
[Per Wikipedia – APPs website did not come up]
The Progressive Populist is an independent newspaper that reports from the Heartland of America on issues of interest to workers, small business people and family farmers and ranchers.
Politicians who do or offer to do what a plurality or majority of people SAY they want rather than what they would probably want were they well informed.
For instance, mandating that all wages in the country are doubled as of now. Or interfering in the Steven Avery case simply because of a popular documentary on Netflix.
I’m adding another voice in favor of populism: you don’t get to sneak around changing the definitions of “well-informed opinion” and “rationality” to automatically privilege elite opinions over those of the masses! Opinions are only right or wrong by virtue of their entanglement with facts, not because of who made them up.
I didn’t find anything at all about Trump on Scott Aaronson’s blog.
Scott Adams. There are too many Scott As.
Yes, this must be settled…who will win in a three-way fight to the death?
It’s worth noting that Adams _says_ things about having 99% certainty, but certainly does not act like it. He’s more than once given himself fairly explicit outs: his blog is “entertainment”, it’s not a mathematical prediction but a “model”, etc.
Given that he is going to welch on how certain he has been all this time (I am 99% certain and will bet with any independent counterparty that Scott Adams backpedals on how certain he was that Trump wins, conditioned on Trump not winning), I don’t think it’s meaningful to claim he is in fact certain now. He’s lying, or at best choosing to not understand what certainty _means_.
Yeah it’s weird to me how many people get confused about Adams’s 99% credence. It’s like they haven’t read the Parable of the Dagger.
Who doesn’t like populism?
Who are the elites? Obama/Clinton/Sanders are certainly populist when they want to be and the media doesn’t seem to mind.
Most people in the U.S. like populism in the abstract or in generalized rhetoric; it seems democratic and in line with our anti-aristocratic tradition.
The problems come in the concrete, when we start having specific policy proposals based on lowest-common-denominator pandering to the masses. That’s less popular, especially among elites.
Very thoroughly disagreed. “Populist” is a label the VSP elite throws at whatever it doesn’t like. Sometimes it doesn’t like Donald Trump because he’s a fucking fascist, but also sometimes it doesn’t like “The Rent is Too Damn High” Party because they’re legitimately pointing out the class divisions in society.
All the non-people.
There tend to be a lot of them, actually.
As a result, I think Trump – who I acknowledge is a competent persuader – will still be sunk by the fact that most people genuinely don’t like him and associate him with toxic ideas like racism and populism, plus he has little governing experience.
It’s 2016. Nobody cares about being called a racist or a populist now. 😉
That said, I don’t cardinally disagree. I myself put a Hillary Clinton victory at 70%, regardless of my own unconcealed loathing for her. Of course some things can always emerge to sink her e.g. the emails, Benghazi, but it seems that the US deep state has her covered there.
> Nobody cares about being called a racist or a populist now.
Trump certainly doesn’t care what people call him, nor do his supporters.
But out of the voters who are not yet his supporters, are there enough left who don’t care that he can win the support of half the electorate?
I think Trump has a better chance of beating Hillary than any of the other Republican candidates. I don’t see Cruz winning any more states than Romney did, and I don’t see Rubio getting people any more excited than Romney did.
However, Trump’s main advantage is that he won’t be able to be Rick Lazio-d. You know, the hapless guy who ran against Hillary in her first Senate race. Had the audacity to run against a woman, you see.
If anyone other than Trump wins the nomination, this will happen:
Anyone But Trump: “I think I would make a better President than Mrs. Clinton.”
Media: “How DARE you make such a SEXIST remark!”
Anyone But Trump: *whimpers* *apologizes* *weakly scampers away with tail between legs*
Electoral College Map: Bluer than a Smurf orgy.
I paraphrase, of course.
Whereas Trump will cheerfully ignore such criticism and is thus free to campaign with the (metaphorical) gloves off. He’s the only plausible candidate who can do so, and thus is in the best position to hurt Hillary’s chances.
Of course, at this point, I’d put the possibility of a Trump/Bloomberg/Sanders three-way race at over 50%, so what do I know?
I think you’re right. As unlikely as this seemed to me six months ago, Trump probably has the best chances of beating Hillary, precisely because she has so many built-in advantages that she essentially wins by default. To have a reasonable chance of beating her requires someone who won’t play by the rules. Trump is one of my last choices among the potential Republican nominees, but if it’s him or Hillary I’d probably vote for him.
I think looking back at the 2008 Democratic primary provides some evidence that this narrative is really unrealistic. Proxies for Obama and Clinton accused the other side of sexism/racism, but nobody turned tail and did the beta-moonwalk you describe.
And the same is true of the general election in ’08, actually. (I guess McCain’s “Obama isn’t an Arab” moment could be called a turning tail of this kind – IF YOU’RE TOTALLY DELUSIONAL.) McCain and Palin never turned tail and cowered for any of the Obama-socialist stuff, and Obama-Biden never turned tail and cowered for any of the Palin-fluffbrain stuff.
The fact that they could accuse each other meant that they didn’t have to turn their tail and whimper. White male candidates don’t have so many options.
“…the beta-moonwalk you describe. ”
Strutting about like the cuck of the walk?
My value system defaults to lefty, but I’m still strongly considering voting Trump if it comes down to Trump vs. Clinton. For this election, insider vs. outsider is a more important axis to me than left vs. right.
My only hesitation there is that I think Trump may be exaggerating the degree to which he is an outsider.
“For this election, insider vs. outsider is a more important axis to me than left vs. right.
My only hesitation there is that I think Trump may be exaggerating the degree to which he is an outsider.”
These are my feelings almost exactly, though I’m more right-wing.
I’m so unhappy with the political status quo that I’d rather elect Mickey Mouse or Vermin Supreme than Hillary or Jeb! or any other member of the current establishment. And I also agree that my biggest fear about Trump is not that he’ll be really crazy (though I think there is some small chance of that), but that he won’t do anything to rock the boat.
I’d vote for Trump over Clinton, but Sanders over Trump (and I’m very much a righty). Sanders won’t have a cooperative congress to do any damage with and I doubt he’ll be as aggressive with the executive orders (though I could be wrong).
Everyone talks as though Hillary is this Nixon level political genius who can ram through legislation, in reality her political career didn’t even take off until she ran for the Senate in one of the most liberal states in the country after her husband was president. Similarly she lost the election to a neophyte named Barack Hussein Obama.
“Rubio could do okay in a general”
That’s an odd prediction. Your P(Rubio Nominee) is .4 and your P(Clinton president) is .6 and you P(Trump Nominee) is .6 and your P(Trump president|Trump Nominee) is .2 which means your P(Rubio President| Rubio Nominee) is 70%.
Rubio with 70% chance of winning a general is a bizarre prediction. Are you aware that he wants to prohibit abortion even after Rape and when the mother’s life is in danger and when the fetus has horrible defects? And that there is lots of tape of him announcing that on teevee? And that CO is the key swing state in the past two elections and NV has a legal marijuana initiative polling very well and is also a key swing state and that Rubio has promised to prosecute users and neighborhood stores and gardeners and shut down marijuana in those states?
I put P(Rubio President|Rubio Nominee) at about 15%. P(Rubio wins CO|Rubio Nominee) I put at 3%. Rubio is just extraordinarily badly positioned to make any Electoral College progress. Bush or Kasich or even Christie would be more viable choices from the BARACK wing of the R party.
I don’t know where people are getting these very high probabilities of Democratic wins from. The ruling party loses more often than not after 8 years in power and general election polls have Clinton behind any non Trump candidate currently. I find it hard to see the Republicans as anything but marginal favourites at this stage.
When a party has held the White House for 8 years, they have historically continued to hold it about 50% of the time (8 out of 15 cases).
You’re reaching too far back to get that n=15. The past is a different country.
Since WWII, which is a better mark of “modern” America, there has been a clear pattern of 8-years-and-switch.
Since WWII, candidates who ran after at least 8 years of rule by their party include Truman, Nixon, Humphrey, Ford, Bush I, Gore, and McCain. Of those 7, two were elected (Truman and Bush I) and one more won the popular vote (Gore). With that sample size I don’t think there’s nearly enough evidence to call switching a “clear pattern”.
EDIT: I guess by counting Bush I twice you can tilt the numbers a bit towards switching but the sample size issues remain.
Truman’s term bracketed the end of WWII – I did not specify the end of WWII, sorry.
And really, you invalidate your point by the “Gore won the popular vote!” line. It’s not relevant.
The only president to buck the 3 term limit trend since the end of WWII is Bush. This is absolutely a pattern. The electorate tolerates one party jacking things up for a while, then calls for the other side to give it a go.
@Nathan: people get very high estimates of a Democratic win from the same place Pauline Kael did: the typical-mind fallacy. “If I can’t imagine anyone supporting Trump/Rubio/whomever, then obviously that candidate has no shot!” Would that it were so.
I don’t totally buy your political metronome theory, either. I tend to take my naive estimate of my least-preferred candidate winning, then double it; this is my extremely rough heuristic for tacking against my own bias. I come up with P(Rubio President|Rubio nominee)=.4, P(Trump President|Trump nominee)=.6. Neither is a courageous forecast!
I’m not saying that a long time in power guarantees a loss for a party, just makes it marginally more likely. I admit to not doing any sort of statistical analysis of this but from observation it seems to be the case in general (though a few countries like Singapore are a decent counter example).
The electorate continues to shift towards an increasingly pro-Democrat composition. It becomes 2 percentage points less white every election.
The parties can pivot and maneuver around this, but the default position of every election is that the electorate, by its identity and nature, is more hostile to the Republicans than it was during the previous cycle. And the Republicans have lost the popular vote in 5 out of the last 6 Presidential elections, so the baseline assumption should be that they lose this election as well.
I’m inclined to think that there will not be another Republican president for a generation, or perhaps ever. At some point, a new party might form, or it might not, and the Democrats will become the “party of power”, in the mold of Mexico’s PRI or Japan’s LDP.
One can just as easily make the statement that Republicans have won the presidency 5 of the last 9 elections. Before Obama we could have said Republicans had won it 5 of 7 elections.
Picking arbitrary cut off points makes for good talking points, but not necessarily good analysis.
I don’t disagree that the demographic trends do not favor the current Republican coalition. Eventually Republicans are going to need to do something to change how their coalition is made up or risk becoming a semi-permanent opposition party.
Besides, Blacks are not growing faster than whites, and Latinos start identifying as white a few generations in.
Even the guy who postulated it first ended up disavowing the “Emergent Demographic Democratic Advantage” (or whatever it was called).
Respect to SA for acknowledging the silliness of some Rubio predictions without trying to go back and change them.
Probably a bit late posting here, but four points:
1. I’ll agree that cut-off-type arguments are a bit arbitrary and serve more for rhetorical effect. But you have to go much further back to pick a cutoff that looks favorable to the Republicans. In particular, you have to go back to an era where the U.S. electorate was near its historical norm (80-85% white) as opposed to the current distribution (70% white and dropping 2 points per cycle).
2. Blacks ARE growing faster that whites:
e.g., 75.5% white to 11.6% black in 2004. 71.1% white to 12.0% black in 2012. Moreover, the global population of blacks is exploding and eager to migrate to the U.S., while the global population of whites dwindles and is, on average, much less eager to migrate.
3. Is there any evidence that third+ generation Latinos identify as white and vote like whites?
From what I’ve seen, second+ generation Cubans tend to turn Democrat in contrast to their first-generation Republican parents. Mexicans start Democrat and stay Democrat, though the Mexicans in e.g. Texas lean more Republican than those in California, for race and class reasons.
4. The Emerging Democratic Majority writers were too early, and now they’ve capitulated too early. It doesn’t make the argument wrong, merely its timing. They thought the Democrats would retain more of the white working class than they did. It doesn’t matter — the white working class is declining in relevance nonetheless.
I hope you’re saving all these in a file so that when (not if) the Trump landslide happens, you can make a stunning mea culpa post analyzing your own bias that will receive over 100k hits, just in time for the end of 2016. This makes the most sense to me, because you said “most people genuinely don’t like him and associate him with toxic ideas like racism and populism, plus he has little governing experience” despite *knowing* that “most people” aren’t SWPLs, and “As for Hillary, basically I just can’t see who else besides her would win.” despite *knowing* (I hope?) that the indictments are coming.
But will you be too polite to analyze the politically-motivated self-deception of “rationalists” in general over it? When, not if, the Trump landslide has already happened, of course. The rationalization scramble will be in full effect then.
There’s not a 99% of a Trump landslide, though. There’s a 99.99% or so chance of (Trump Landslide ∨ Trump is assassinated before Election Day). It’s not fair to ding Scott Adams for his deliberately simplified model that doesn’t include things like assassinations.
According to the prediction markets, Trump has a 44.9% chance of winning the primaries and a 17.5% chance of winning the election (thus, 39% chance of winning conditional on winning the primaries). If you think a Trump landslide is inevitable, then there is a ton of free money going right now while you’re wasting time arguing with people on the internet.
there is a ton of free money going right now while you’re wasting time arguing with people on the internet
A “ton” of free money would be approximately one million US $1 bills. But let’s say you were exaggerating a bit. Where on the internet is there even a hundred thousand dollars in free money available for law-abiding US citizens with a solid understanding of electoral politics?
As has been repeatedly pointed out on this very thread, no there isn’t. Prediction markets and bookmakers are divided into two categories: those that pay only tiny sums of money, and those that American citizens are liable to be arrested if they are caught dealing with. Though realistically, the Feds would probably just seize the money and forget about the arrest, so long as you don’t make a fuss.
I would be in favor of an immediate banning for any idiot who insists on repeating this annoying fallacy.
“tiny sums of money” for a given value of tiny. I made 150 bucks on PredictIt through arbitrage, and people actually willing to hold real positions could probably make a lot more. For example, the market grossly overweights Bernie Sanders’ chances of winning.
Not enough to quit your day job, but 150 bucks is several hours’ work for many people.
“despite *knowing* that “most people” aren’t SWPLs”
Most people aren’t just not-SWPLs, they’re not even close to participating in a conversation wherein that statement isn’t gibberish.
The irony of the alt-right and its variants is that they correctly intuit that illiberal leftists are a tiny minority whose voice is vastly out of proportion to their numbers, probably because they’re a bunch of underemployed internet addicts, but somehow think that the reaction against those same people is some sort of overwhelming tide. It’s not. Most people live normal lives and spend their energy on their jobs and their families. They don’t know about weirdo-on-weirdo online political fights and they wouldn’t want anything to do with them if they did …and most of those people think Donald Trump is an obnoxious jerk, not so much because he’s a racist but because he’s rich, arrogant, and unpleasant.
What is SWPL?
“What is SWPL?”
“Stuff white people like,” named after the blog. Young blue-tribe urban yuppies. Probably could (and, really, probably should) be differentiated from illiberal leftists/”SJWs” who are a smaller, more radical, and more online-oriented cohort, but I think that comment was treating them as synonyms and I did the same.
What nil said. SWPL has become shorthand for the WP that the blog in question meant when it described things that WP allegedly like. “Whole Foods.” “Mountain Bikes.” Etc.
Do not, by the way, under any circumstances Google the phrase “stuff black people like.” Or, even more importantly, “stuff black people don’t like.” At least not on a work computer.
I don’t *know* that indictments are coming against Clinton.
Assuming for the moment that I think she rightly deserves them for (say) her handling of her email server, there is no way that she gets indicted while being politically successful.
If her campaign *fails*, she might find herself under investigation, because everyone loves to kick a loser while they’re down. But people in government don’t pick fights with those in power, or those ascendant to power.
Say for this?
One notes that political success being a protection against prosecution was exactly the last nail in the coffin of the Roman Republic. Julius Caesar crossed the Rubicon, got to Rome to declare himself a candidate, and henceforth had to succeed or be prosecuted.
The real question is whether Obama wants to destroy Hillary or not. If he chooses, he can make the law apply to her, throw her in jail, and sink her campaign. It’s not going to happen without his tacit approval.
I would be very surprised if he did so. But as Professor Seldon pointed out, it is much easier to predict the actions of large groups than of individuals.
it wasn’t that success protected him, it was that his protection wore out when he was out of office. The senate refused to let him run for office again (which was technically not allowed but exceptions were made as a matter of course).
I’m well aware most people aren’t SWPLs. But the polls show independents and Democrats hate Trump. Right now his only hope is that he can pull off an amazing pivot and be like “Haha, only joking, I’m actually a pretty responsible respectable guy”. And I think he will try, and do it well, but that most people’s opinions of him are already too fixed for it to be enough.
Happy to make a monetary bet with you based on your 99.99% odds (including assassination), or something logarithmically halfway between that and my 20% odds.
The thing I wonder about is not Trump pivoting to pick up the middle and independents, but pivoting to pick up the less educated left.
I could see Trump doing well with African Americans and white Blue Collar Dems. Trumps “racism” has been targeted mexicans and muslims and I am not sure that that bothers African Americans that much.
Trump might be able to put together a coalition of folks without college education on both sides and die hard republicans who can’t bear to vote for Hillary. Is that enough to form a winning coalition. I don’t know.
If eccdogg is right and Trump achieves “wider respectability” in the sense of attracting dissatisfied Democratic voters, but without pivoting on issues such that he gains “wider respectability” among highly educated professionals… then was Prediction #22 right or wrong? I know it has an asterisk, but this seems not so much “subjective” as “I don’t actually know how Scott himself would score that.”
I’d be willing to bet that if Trump is the R nominee, and that either Clinton, Biden, or Sanders is the D nominee, that Trump outperforms the R share of the vote among Hispanics *and* blacks. Specifically, that his national popular vote share of the black vote and of the Hispanic vote is higher than the largest national popular vote share that any of Bush 2000, Bush 2004, McCain 2008 and Romney 2012 achieved.
Yeah, but none of this matters, because blacks are only 12% of the population and they vote 90% Democratic. Even if Trump can triple the usual Republican share of blacks, that only gains him about 2-3 percentage points. His disadvantage among Hispanics (even for a Republican) will be many times that, and then we get to the part where he fails among independent whites.
Eh, let’s not forget about half the voters being women. The most hilarous part of modern democracy is that it has a potentially big seduction element. Seems to have worked for Bill Clinton and Berlusconi. This could be more decisive than say a bunch of white male Dem union workers.
This is why Hillary cannot win. Women voting for a woman? No way. Crab bucket effect. Women are jealous. She can get the college feminist vote and that’s it. Men voting for an old, unattractive woman? Like hell. Women voting for an aging charmeur with an arrogant vibe? Hell yes. Men voting for an aging charmeur with an arrogant vibe? Not really sure, men can be jealous too, but far more likely than voting for an old woman.
> This is why Hillary cannot win. Women voting for a woman? No way. Crab bucket effect. Women are jealous. She can get the college feminist vote and that’s it. Men voting for an old, unattractive woman?
…ahem, we went and elected and then re-elected an old, unattractive woman as our president, though some time ago. Until recently, Frau Merkel did very well in Germany for a quite long time.
Maybe women in the States are very much different than in N-W Europe. Cultures differ. However, reasoning based on such crude stereotypes feels very …out of place.
Attractive trait on a female leader is OP, +30 attraction trait to male vassals along with the ability to invite male counselors and claimants from all over the map.
it’s not that, although I do think Americans are more over-sexualized as a culture than Europeans, selling infomercials with tits or deep male voices works better. But I think it is rather that anyone could have won against Gerd Schröder and the likes. Remember the situation – high unemployment, a rather dysfunctional looking economy, and Schröder harping about the rich being too richy like an utter dolt, I think even for people who usually have leftie economic views it was clear it is not the right time for that, this time is about growing the economy, not redistributing it. And then a woman stands up, looking like an Ossi Thatcher, representing a party that was still considered kind of conservative back then, and says my No. 1 goal will be cutting unemployment and judge me only on that and nothing else. That just had to win, in that kind of situation.
I don’t see a Trump landslide. It depends on who the Republicans can put up against him. If he does win the party nomination, it will probably be him versus Hilary and I think she’d just about win (unless the Democrats do the usual internal throat-cutting and all the Bernie Bunch flat-out refuse to vote for anyone but their messiah).
The main problem on the Republican side isn’t so much that Trump is beating them, it’s that of their proposed candidates so far, nobody stands out (except in a “Oh God, do we want ‘watching paint dry guy’ or ‘wet cardboard box guy’?” manner) as the kind of candidate who can mobilise support across the entire country.
Trump is the Reality TV Show candidate and I think people are giving him support in much the same way they’d vote for the most outrageous character on such a show; part of it is enjoying the reactions of the offended (and really, when transgression has been made the highest value, it’s amusing, even where I dislike Trump, to see those who’d fervently defend offending the majority opinion as a duty to attack sacred cows suddenly clutching their pearls over blasphemous rhetoric). When it comes down to really voting for a president, though, I don’t know if he can get serious votes.
“I think she’d just about win (unless the Democrats do the usual internal throat-cutting and all the Bernie Bunch flat-out refuse to vote for anyone but their messiah).”
That pretty much happened in 2000. A sizable number of Democrats “voted their hopes not their fears” in November for third party Nader, with the result of electing Republican Bush II.
I hope Sanders won’t run as third party and split the November vote. I hope the worst danger is that some DINO Democrat will slip in and get the Democratic nomination. (Clintonista here.)
Scott Adams says a lot of weird stuff. I used to think he was just screwing around and throwing ideas at the wall to see what sticks until I saw the extent to which he seriously defends his ideas about gender politics in the face of heavy criticism. I think he’s just kind of crazy. (Also a huge misogynist, but that’s a debate for another day.)
When even otherwise pretty liberal guys, like Scott A and Scott A (and also Scott A?) say that there’s a *HUGE* problem with what feminism has done to relations between the sexes, perhaps there’s really a problem with feminism? Even if the problem is that it’s turning likely allies into misogynists, that’s still a big problem for feminism.
Or maybe reality is misogynist.
Scott A certainly isn’t liberal in any sense, and I don’t trust his opinion on anything except making funny cartoons. Scott A has had negative personal experiences with feminism, but by itself that doesn’t mean there is anything wrong with the movement as a whole, unless it can be proven that his experiences generalize. Scott A’s whole shebang is criticising his ingroup. Since no political movements are perfect, it is unsurprising he criticises feminism occasionally. That doesn’t mean Scott A is right about feminism (or indeed anything other than funny cartoons).
Objection: Scott A(dams) holds many reasonably liberal views. (He is against the drug war, he is very strongly pro-euthanasia rights, I’ve never heard him indicate that he was an anti-abortion advocate, etc.) He is certainly not a conservative, in the sense of “conservative” being the opposite of “liberal.”
Overruled – you’re not supposed to indicate which Scott A you’re talking about. More seriously, Scott Adams seems to be a typical internet libertarian with a few odd opinions. As with most members of that tribe, he has many Blue-tribe views (more than I expected, from a quick read of his wiki page, I wouldn’t have expected him to be vegetarian) but also a few decidedly non-Blue-tribe ones.
You’re probably right, but isn’t it so very depressing that Sanders would do far, far more to help minorities than Clinton ever will?
Perhaps one might contrast ‘might hope to’ from ‘would succeed at’.
If there is one thing I could achieve by being made Empress of the Galaxy, it would be to banish, destroy and otherwise extirpate the notion that any one politician of any one party, no matter their principles, good intentions, or high-mindedness, is going to be the Messiah. (My second diktat would be: “Campaign promises are not to be believed, and if you vote in the expectation because of them that you will get a pony and a lollipop when the candidate is elected, and you are over the age of four, you deserve all you get – which will not, in fact, be a pony and a lollipop”).
“X would/could do so much more to help the endangered purple-furred sea kitten! (if only Y were not opposing them/the evil meanies in the opposition were not blocking them/they were given absolute power/we all join hands and wish really hard)” is the kind of thing that makes me sigh heavily.
I’ve seen the crusader politicians on this side of the water. I’ve seen the non-politician candidates going in as independents in their white suits (literally in one case) elected on a wave of popular support to reform and uplift, and I’ve seen them serve out their time and retire or petulantly give up, with little to none of the promises they made being fulfilled, because the system and the structure and the inertia and the necessity of ordinary compromise and deal-brokering and what power can be exercised how is against that type of root-and-branch reformation.
Sanders might or might not be all that you hope. But that he would achieve so much more for minorities is, I submit, a pipe dream. He could achieve something. But that he’d be able to implement all his campaign promises as-is, no changes, re-tooling, dropping certain elements? No, not at all.
For goodness’ sakes, have you not learned from the Obama campaigns? When he won his first term, and there were frankly embarrassing paeans to him as the star-destined enlightened ruler that would lower sea levels and make us all live in Utopia?
And when the promised castles in the air for all did not eventuate, well, it was because of racism and those nasty meanies in the opposition blocking every good, pure thing out of simple evilness.
I’d really like to see an impartial, unbiased “campaign promises beforehand/actual implementation afterwards” list of what was put forward as policies for both his campaigns done by someone, but I don’t think I’ll get it: on the one hand, there would be too much “And he lied about this!” and on the other hand “And the Evil Meanies tied him up and held him prisoner while they rammed through their evil amendments to the Bill!” stuff.
Does nobody remember the carefully staged PR shots of “Our Take-Charge Kick-Ass President and his team overseeing the defeat of the wicked bin Laden”, which if a Republican had done it (as Bush did with his ‘put on a flight suit and get the photo taken with serving military’ which was every bit as staged a stunt for public consumption), would have been excoriated as populist pandering to base revenge instincts?
That’s the pragmatic politics in action, where it’s been decided by the special advisers and the focus groups that it would really help your image, Mr/Ms President/Prime Minister, to be shown as decisive and active not reactive, and the party needs a bounce in the polls for the forthcoming local elections. I’m not quibbling with that, he’s a professional politician, he knows the score.
But wistful hopes that Our Guy (or Gal) is bigger than the system – no, I can’t subscribe to those anymore.
“But wistful hopes that Our Guy (or Gal) is bigger than the system – no, I can’t subscribe to those anymore.”
Perhaps the worst case is when the candidate believes that himself. (Which the Clintons, having been in the White House for eight years, are unlikely to believe.)
Campaign promises: yes because practically the whole world uses the free mandate system where the law explicitly says promises are not binding and reps can rep as they wish. Mandates could be bound / imperative…
In case anyone wants more on the subject:
Of course, if you were Empress of the Galaxy, you’d *still* have to get your proposal past an extra two or three layers of government.
Two sf questions: There’s a rant by an Empress in one of van Vogt’s Isher novels about how she can’t get people to do what she wants because every order she gives has to be interpreted. Does anyone happen to have the rant handy?
Does anyone know of detailed speculation (whether in fiction or not) about how organization might be done in really large populations and/or with significant time lags? My guess is that it won’t be all that much like feudalism.
“I’d really like to see an impartial, unbiased “campaign promises beforehand/actual implementation afterwards” list of what was put forward as policies for both his campaigns done by someone, but I don’t think I’ll get it”
For Obama, we sort of have this.
Politifact tracked Obama’s campaign promises and subsequent actions.
His current score is 45% kept, 25% compromises, 22% broken, 6% in the works, and 2% stalled.
It’s not perfect — a lot of the promises they score as kept I’d score as compromises, and a lot that they score as compromises I’d score as broken. And of course, the stalled and in the works promises can basically be considered broken at this point.
But they do manage to mostly avoid “he lied” and “it’s the fault of those meenies.”
Of course, there is also http://www.politifactbias.com/ for a second opinion on politifact.
@ Nancy Lebowitz
Thanks for an interesting link, though imo the writer’s take is much over-simplified.
The last time he ran for President, Mike Huckabee described an approach … don’t remember what he called it … of inventing new solutions that were so practical that both sides could accept them. When the Clintons in the 90s did that, it was called ‘Third Way’.
A President has resources to find such solutions, and to sell the public on them. (And, given the intelligence and the will, to work out compromises with various Congress Members, and if necessary to keep vetoing larger bills till Congress gets tired of trying to over-ride the vetos and incorporates his solution.)
Significantly, I think the other Scott A wildly overestimates how many people listen to the candidate they’re voting for.
Even if Trump could get the vote of anyone who heard him speak for >5 minutes, he still couldn’t win by the kind of margins Scott A has occasionally implied. 20-30% of the voter pool is lost purely by having an (R) or (D) next to your name. And, along with the party-line faction, there are a lot of people who couldn’t align their self-image with a Trump vote under any circumstance.
My estimates are a bit more Trump-favoring than yours (I think skipping the next debate will be a huge win for him), but I still think ScottA2 badly misjudges how many people are immutably opposed to Trump.
But your reason for Sanders being unpopular among minorities is that minorities are less eager to get behind non-mainstream-seeming things. Isn’t it possible that after Sanders won Iowa and NH, he would seem mainstream, and minorities would get behind him?
It might be very well be true that all your friends think that racism and populism are toxic ideas. I don’t think that a majority of US citizens thinks that way.
Most of the electorate follows also the news a lot less we do.
I hope nobody here takes Scott Adams at all seriously. I generally try to give people the benefit of the doubt, but his blog is almost entirely composed of obfuscation, misrepresentation, and other things that would generally be considered ‘dark arts’ by rationalist-types.
He also seems to have absolutely zero interest in actually defending his supposed point of view from reasonable criticism. I posted a detailed criticism of his claims of being an infallible political predictor, and he failed to respond, despite the fact that his own readers voted it to the top of the comment thread. He is frequently active in comment threads, and I posted my criticism a number of times, so I assume that he just had no interest in engaging.
Adams may in fact have some talent for political prediction, but his constant stream of half-truths makes it hard to actually tell. He’s clearly not interested in actually making fair predictions. He’s working against the values of honest, clear communication and reasonable discussion, and he’s trying to set himself as some sort of political prophet in the process. I’ve given up on getting Adams to clarify or defend his actual predictions, but I think it’s important to make sure that he pays some reputation cost for his deception.
If anyone cares, I can re-post my criticism of Adams’s ‘predictions’. They may not be up to date, since I decided that not reading his blog was a better choice for my overall stress level.
Please do repost them? I’d be interested – I look at his blog periodically, and it seems to me he goes way too far, but in a way that overcorrects for the General Standard Opinion.
I think this is the most recent version (including corrections and suggestions from some other readers) of my response to Adams’s post claiming to have a perfect prediction record in 2015 (http://blog.dilbert.com/post/136261193951/ranking-the-best-political-pundits-of-2015). The gist of it is that his actual predictions aren’t too bad, but he exaggerates his success, misrepresents the circumstances of the predictions, includes trivial predictions, and conveniently ignores predictions when it suits him. If you think I’m getting anything wrong, or being too uncharitable as a result of my general dislike of Adams’s tactics, please let me know.
1) Trump has gained some popularity, but he hasn’t won anything yet. I give Adams a lot of credit for his prediction of Trump’s longevity, but he didn’t just predict that Trump would gain popularity, or even that he would win the nomination. He predicted that he would win the general election in a landslide. He predicts that Trump will get 65% of the vote, although he does not specify whether he’s referring to the popular or electoral vote. Even if Trump wins the nomination, and easily wins the presidency, Adams could still be wrong. Obviously, if he turns out to be correct, that’s an extremely impressive prediction. (http://blog.dilbert.com/post/131552504961/trumps-third-act-part-of-the-trump-persuasion)
2) Bush’s poll numbers were slowly dropping before Adams said anything about ‘low energy’ (http://blog.dilbert.com/post/127715904536/trump-persuasion-alert-the-bush-slayer-comment) They have continued to drop, but the RealClearPolitics average doesn’t show any particularly dramatic change.
3) Adams claims that Fiorina self-immolated in her debate performance (http://blog.dilbert.com/post/129289732811/carly-fiorina-and-the-wizard-filter). My reading of the poll numbers show that Fiorina was consistently polling in the mid-single digits, had a big boost for about a week as a result of the debate, and then returned to the mid-single digits, and slowly dropped a bit lower.
4) Adams gets some credit for Carson, but he did not actually make a prediction here. The blog post in question (http://blog.dilbert.com/post/133133670296/trump-on-carson) praises Trump’s speech, but it does not make any prediction about the fate of Carson.
5) To his credit, Adams made a specific prediction here (http://blog.dilbert.com/post/135152311211/calling-the-clinton-top-trump-persuasion-series) saying that this was the high point of Clinton’s poll numbers in a matchup against Trump. According to Wikipedia (https://en.wikipedia.org/wiki/Nationwide_opinion_polling_for_the_United_States_presidential_election,_2016) there have been seven polls since this prediction on 12/13. I list the Trump/Clinton margins here, compared to the previous margin from each pollster:
Rasmussen – went from Trump +2 in October to Clinton +1 in late December
CNN/ORC – went from Clinton +3 in late November to Clinton +2 in late December
Ipsos/Reuters – Clinton +12 in late December (no previous data)
Emerson College – went from Trump +2 in October to Clinton +2 in late December
Quinnipiac – went from Clinton +6 in late November to Clinton +7 in late December
Fox News – went from Trump +5 in November to Clinton +11 in late December
Public Policy Polling – went from Clinton +1 in November to Clinton +3 in late December
Everyone can make of this what they want, but I’m not quite seeing how Clinton’s margins over Trump are slipping.
6) This looks like a correct prediction from Adams (although I was unaware that Trump had actually stated an intent to improve his likability, which suggests that this was not a terribly difficult call). However, I notice that this post (http://blog.dilbert.com/post/131997255161/trumps-town-hall-performance-give-him-a-grade) contains a second prediction by Adams. He predicts that Trump’s likability will sharply increase within the next month (after 10/26). I wonder if someone with access to good polling data can analyze this claim? I could only find one poll, from 12/11, that had Trump’s likability at 24%, the lowest of all the candidates surveyed. However, I couldn’t find any older data to do a comparison. The closest I could find were favorability ratings, which have trended slightly toward unfavorable over the last couple months.
7) Adams mentions the ‘President Trump’ thing here (http://blog.dilbert.com/post/135152311211/calling-the-clinton-top-trump-persuasion-series) after it has already happened. He was correct in that it did happen again, but that’s an observation much more than it is a prediction.
8) Adams correctly predicts that he will be ignored by the mainstream media. Given that he is making outlandish predictions (65+% of the popular vote goes to Trump), claiming a 60% chance that he can stop Trump from being elected (http://blog.dilbert.com/post/134861704021/my-offer-to-stop-donald-trump) posting sexual hypnosis stories (http://blog.dilbert.com/post/136330131311/hypnotizing-you-to-have-the-best-new-years-day) and generally not behaving like someone making serious predictions, that’s not exactly a shock.
9) He also correctly predicts that there will be competing explanations for Trump’s continued success. This seems very obvious to me, but maybe I’m being uncharitable here.
“He also seems to have absolutely zero interest in actually defending his supposed point of view from reasonable criticism. ”
That’s not Reasonable Criticism. That’s just trying to get laid by his Moist Robot theory.
Which only makes me quote Edmund Burke: “He that accuses all mankind of corruption ought to remember that he is sure to convict only one.” I’m willing to believe that HE’s a moist robot. (Which is more than he is. He talks about programming other moist robots as if he could choose to do it.)
That’s an excellent quote, and on reading it my mind instantly leapt to Sigmund Freud. It seems as though every “theory of mind” is not only vulnerable to projection, but consists primarily of it.
I’d be curious to see your criticisms too. I have a better view of Scott Adams than you do, but I’m still wary and skeptical.
It appeared to me that several of his big analyses were essentially on the level – he has nowhere near the command of statistics that Scott Alexander does, but he seems to have done his “wage gap” lit review in basically good faith, to the limits of his knowledge.
On the other hand, his “year in review” was self-aggrandizing and absurd – he cherrypicked a dozen predictions out of far more, and then credited himself with several points for vague, fits-all-results suggestions. His Trump posts, which he admits are for entertainment value, are obviously questionable and seem to come from a belief that if Trump is an amazing persuader, then Scott Adams (who has similar teachers) is too.
The most objectionable post I’ve ever seen him write is one promising to use his persuasion on the reader while convincing them that Trump can make Mexico pay for a wall. To his credit, he tagged it with the equivalent of “DARK ARTS INCOMING” – my objection is actually that he failed badly. Whenever he said “and now look, you’re thinking X!” I had either dismissed X or not thought it at all.
I don’t think he’s operating in systematically bad faith (with the caveat that the Trump posts have a warning not to take them seriously) . I think he’s doing some sort of ‘infotainment’ version of rationalism, and badly overestimating his own skill at the process.
Adams’ suggestion techniques will work quite poorly on the average reader of this blog, especially when they are aware that he is trying to use suggestion on them. I had the same experience you did with his “and now you will think X” attempts, but a) I am at least an average reader of this blog, I hope, and b) I’m a hypnotherapist. So I did not think less of him for that.
In my professional opinion as a hypnotherapist, he’s not bad at using suggestion techniques. He’s no Milton Erickson, but then neither are most of the rest of us. While I won’t argue with your other assessments, I wanted to point out that “it didn’t work on me, therefore it doesn’t work/he’s not good at it” is too broad a takeaway.
Its devolved into this
>Trump is good trump is gr8 its all 3D movements braaa u just ain’t seein it
21. Donald Trump will win the Republican nomination: 60%
25. Marco Rubio will not win the Republican nomination: 60%
The combination of the above two predictions seems to imply that you are 100% confident that either Trump or Rubio will win the nomination. Surely that is overstating it, no?
Apparently, there’s an “Everyone Hates Ted” meme going around, where the story is the Republican party elites can’t stand the guy. Which leaves Rubio or Trump as the potential nominees, unless people come back to Jeb Bush or Kasich or Carson?
The Republican party elites hate Trump; if we are nonetheless assigning p>0 to a Trump nomination it seems we ought to do the same w/re Cruz.
There’s a difference between disagreeing with the politics someone stands for and hating them as a person. My understanding is that elites find Trump fairly likable on a personal level and certainly do not detest him anywhere near as much as they do Cruz.
As support for my citing of the meme, Mother Jones: Is Ted Cruz Really an Awful, Terrible Jerk? with quotes from Bob Dole, John McCain, college classmates, Bush administration staffers.
Given Cruz’ poll position, it’s hard to argue that he has a truly minuscule chance of winning.
Beyond that, giving Trump + Rubio a 100% estimate is just bad rationalism – they’re not likely to get caught on video in a three-way with Kim Jong-Un, but it’s >0% and would certainly sink both candidates.
Now, now – a proper rationalist shouldn’t say that anything would certainly sink both candidates. Instead, contemplate the ways that having a nuclear-armed tyrant openly on their team might be helpful to a US presidential candidate 🙂
Trump/Cruz ’16: Nice country you’ve got there; it would be a shame if something happened to it.
And now there’s a New Republic article “Everybody Hates Ted“
That’s a good point. I think it’s because I wrote these over a couple of weeks, and I put the Rubio thing in earlier when he had a good chance, and the Trump thing in later after he was more dominant. If I could edit these, which I don’t think I will, I’d probably change not-Rubio to 70 or 80.
He could assign more than zero probability to Trump and Rubio getting nominated together as co-president candidates.
I hope that there is an Unsong chapter about this.
Scott Adams just came out in favor of a Trump/Sanders co-presidency.
No, he was in favor of a Trump presidency with Trump asking Sanders and Bloomberg for advice. I think Adams is messing around with his own persuasion experiments.
He is rounding into coarse buckets, so T60, R40, C10 could be be correct, eg, T56, R36, C8. (But that’s not what he meant.)
Good catch. I was going to mention the same thing.
>32. Dow Jones will not fall > 10% this year: 70%
Any reason for DJIA instead of SP500 for prediction?
> 36. No exchange of fire over tiny stupid islands: 90%
>44* So-called “Ferguson effect” continues and becomes harder to deny
I’ve been digging around for more data on this – and having a hard time. I’d love to hear more from you on it, unless there’s an extant piece you recommend.
32 – because I’m more familiar with the Dow.
36 – by the four countries mentioned above
44 – Can’t think of anything off the top of my head.
So when you mention “tiny stupid little islands” about “Chinese, Japanese, South Korean, and American” you’re probably talking about the East China Sea EEZ dispute and not the one in the South China Sea (map)?
Dow Jones is already down 9% YTD, so this prediction seems bold.
Unless you mean no 1-day 10% drop, in which case this prediction is incredibly wimpy. (10% drop is like a 1987 crash situation, which has only happened since then during the “Flash Crash”, which instantly reversed itself before the day’s end).
Yeah, this. This prediction is very bold considering where it is today. Unless you mean, 10% further. Either way, very bold considering year-to-year stock market swings.
Surely what he meant was that it would not end 2016 10% lower than it began 2016?
Based on a rough-and-ready look at some Dow numbers I happen to have laying around from 1931 to 1994 (don’t ask me why), I see that such a drop happened 12 years out of 61, or about 20%. So Scott predicting this at only 70% probability is statistically a bit pessimistic.
But given that it is already down 9%, the mean guess for where it will end up should be a bit more than down 9%. Unless you know something I don’t.
@Dr Mist, anon: I read the statement as “drop 10%—from Jan 1 opening price—at any time during the year.” If you read it as “market close on Dec 31 won’t be >10% lower,” of course that’s a much less brave statement.
>44* So-called “Ferguson effect” continues and becomes harder to deny
I’ve been digging around for more data on this – and having a hard time. I’d love to hear more from you on it, unless there’s an extant piece you recommend.
Here is a WSJ opinion piece that name checks the existing sources of the data in support, so you can google those sources directly.
From the other side of the political spectrum, Slate also weighs in.
(I know these links aren’t actual studies, sorry.)
> “Mainstream European position at year’s end is taking migrants was bad idea: 60%”
How does one judge what counts as “mainstream opinion”? Clearly, people at the end of 2016 will still be split, some newspapers will write favorable, heartwarming stories about immigrants and plenty of politicians will cite it as a success. It would be nice if there were a more objective standard one could apply here.
Poll numbers? My default for mainstream is that an idea has plurality or majority support among the population, though then again that puts a lot of ‘things you can’t say in public’ in the mainstream.
If you’re counting on the mainstream media or political establishments to go against immigration your odds are a lot lower than 60%. As for the opinion of ordinary people, even before the New Years gang rapes I would bet anti-immigration positions were commonplace in most of Europe so 60% is likely an underestimate. Either way I’m not sure about Scott’s odds there.
“even before the New Years gang rapes”
As best I can tell, there were no New Years gang rapes. Judged by the accounts I read by victims, the offenses were sexual assault in the form of groping plus theft.
Do you know of any reliable accounts of any rapes as part of what happened in Cologne—I assume that’s what you are referring to?
You are using an outdated definition of rape.
I’m using a modern definition. If I was using the old definition, the stealing would qualify.
I didn’t say you were using an old version. I said you were using an outdated version. 😉
Any penetration against the victim’s will is grounds for rape in the US, if not in Germany.
At the very least you can admit that sticking a finger or foreign object inside is a bit more than is a bit more than a “grope”.
Isn’t non-consensual digital penetration considered rape?
From a German newspaper: “Albers sprach von Sexualdelikten in sehr massiver Form und einer Vergewaltigung.”, which translates to “Albers(the former police president of Cologne) mentioned massive sexual assaults and one rape”.
If you see a cockroach in your kitchen you don’t think “oh well, it’s just one roach” and forget about it. For every one you see there are a hundred that you don’t. It’s time to call somebody.
European governments have in recent years gone to extraordinary lengths to cover up mass rapes by immigrants. The German government did, in fact, try to sweep this under the rug completely. I have no faith whatsoever in the accuracy of any account downplaying this incident.
Although I understand what you’re trying to do with the analogy, try to avoid anything that even gives the appearance of comparing people to cockroaches.
@Scott: this is the problem with political correctness. People will read Dr. Dealgood’s comment and get upset that the analogy might cause offence, and completely ignore the actual problems that it is referring to.
I think that this kind of thinking (“we mustn’t upset or offend anyone! Let’s be all nice and politically correct!”) is precisely the kind of thinking that has landed Europe in this mess, and personally the best outcome I can see from the refugee/terror/rape/barbarity/jihad crisis is that the political left in Europe learns an epistemological lesson about the importance of not conflating virtue with truth.
I remain fairly confused as to exactly why people can’t talk about these things *without* comparing humans to cockroaches. It doesn’t seem like an unreasonable imposition or a condition that would limit the factual propositions that could be discussed, any more than “please don’t break out into swearing-filled tirades at other conversation participants” would.
Even if the only way you could mentally handle the sheer unfairness of the situation was by deciding that the migrants had no moral value, is it really so hard to take something more like egalitarianism as a premise for long enough to make a case under it?
You have to do it in pretty much all walks of life where you want to cooperate with people on matters of mutual interest, even where you’re basically all there out of your own self-interest. And it’s usually more effective to argue within people’s existing values than to try to change them if your concern is just the object level case.
The conclusion people draw is that those people aren’t actually interested in the object level case, but more interested in bringing other people around to their moral value assignment, and are just playing Ethnic Tensions (now with ethnicities!), which naturally gets a hostile response and disengagement. If this is a wild misunderstanding, it’s one that really needs an alternative narrative.
I mean, what you’re saying isn’t that virtue is uncorrelated with truth. If that were the case you can happily filter out all the unvirtuous- you can’t talk to/listen to everyone, you may as well make “be a decently nice person” a filter rather than picking randomly, and either way if you skip over someone who knew something useful you didn’t, you’d encounter someone else who knew it eventually.
What you’re saying is that at least some truths are extremely strongly anti-correlated with virtue- that certain truths will be completely inaccessible if you only talk to nice people. And that this is lasting, as opposed to a temporary hiccup of history and convenience. And these are critical, important truths, and that it’s not just a difference in time to notice them or proclivity to spend their time talking about them. And the minority of people with low standards isn’t enough to let them be introduced to everyone else. All adding up to mean that asking people to play nice does more harm to the discourse than it sets up useful incentive to keep the garden nice.
If that’s true it is very strange, and has interesting implications beyond ‘stop filtering out jerks’- it invalidates the whole “let’s be nice, polite people who find ideas and discuss them internally to arrive at truth” thing as a project. It probably means all forums aiming at rationality need to be almost totally unmoderated. Can one learn these truths on purpose without becoming a complete ass? Does a search for epistemic truth obligate you to try to become one? Should we be putting members of /b/ in charge of running countries? Suddenly “maybe being epistemically rational is not being instrumentally rational” becomes a practical concern. And so on.
> Should we be putting members of /b/ in charge of running countries?
There’s already a movement in America to do that actually, I can’t remember the guy’s name though, I think it rhymes with Ronald Rump?
> All adding up to mean that asking people to play nice does more harm to the discourse than it sets up useful incentive to keep the garden nice.
@JBeshir: there is a tradeoff of sorts between “keeping the garden nice” and being blinded to reality by excessive political correctness.
If we didn’t have political correctness, we would probably miss it; many people would go way too far the other way and express more xenophobia and racism than is optimal.
I think the lesson that Europe is learning is that the optimal amount of xenophobia is not zero.
Xenophobia protects us from the genuine, real harms that people from foreign cultures can inflict on us, especially when our political leaders give them a blank cheque to do whatever the hell they want and anyone who challenges them is a racist, or anyone who challenges them is “helping the far right” etc.
My metaphor, if it’s read like a piece of prose English rather than mined for keywords, is pretty clear that the rapes are the cockroaches here. Otherwise it doesn’t make any logical sense: nobody is covering up the fact that Muslim immigrants exist, at least as far as I know of, but I rather explicitly stated my position that this sort of mass rape is in fact largely hidden from view.
Anyway, I’d ask that if you don’t want to talk about that issue (understandably, I don’t even like thinking about it) that you don’t drag the conversation down weird tangents but rather ignore the thread entirely. There is a hide button under every post so there’s no need to derail.
(Not accusing Scott of derailing btw. He obviously got the metaphor even if he thought the phrasing was poor, and I will respect his wishes in that regard going forward.)
I think optimistically, people might learn that there’s some flaws in their current conception of virtue, and learn something like “culture is a thing, and people who care about everyone’s well-being should be interested in it, and it’s inappropriate to attack people for being so”. Not xenophobia in all its connotations but culture as a valid area of concern.
Ideally we’d even see this extend to a general demand for presumption of cooperation when it comes to debates of fact, statistical probabilities about the people who like to talk about a given topic to be set aside. That’s wildly optimistic, though. In either case it’d be a trade-off, in that you’re throwing away evidence you could be using to detect defection, but not a big one.
I don’t know how much I expect that to happen, as opposed to temporary pressure resulting in short term actions but no further changes, or people shifting politically to decide that selfish ethnic identity-based politics is actually a good idea. I’ve been pessimistic, but the reactions to Cologne have been surprisingly promising in this regard so far.
Edit: @Dr Dealgood Yeah, okay. I didn’t have any strong feelings about your original post, just responding on, as you put it, a weird tangent. Although I guess I’ve looped back here to some theorising of what might or should come of all of it in the longer run so I’ll leave this post.
It isn’t comparing people to cockroaches. It’s comparing incidents to cockroaches.
And the problem is that I can’t just say “okay, I’ll never use a cockroach analogy again” and be okay forever. There are a whole bunch of arbitrary things I’m not allowed to use, of which this is only one example. The cumulative effect of all these arbitrary prohibitions is to make discourse almost impossible. It’s like the gun restrictions or abortion restrictions mentioned in other threads where guns./abortion can’t be directly prohibited, so their opponents try to make them die from a thousand cuts.
(And the rule “you can’t say something that even gives the appearance of comparing people to cockroaches” is much broader than just “you can’t compare people to cockroaches”, and the chilling effect of the former rule is much greater. Also, you’ve created a rule which rewards misunderstanding and creates disincentives to understanding, since someone who misunderstands you can shut you up. Creating a rule which rewards X will lead to more X.)
“Xenophobia protects us from the genuine, real harms that people from foreign cultures can inflict on us, ”
There are children with no xenophobia, fears of strangers, or racism.
They suffer from Williams syndrome and must be carefully taught to consciously think that if they don’t know a person, he might not wish them well.
It isn’t comparing people to cockroaches. It’s comparing incidents to cockroaches.
And you don’t see how “incidents” and “cockroaches” are such categorically different things that miscommunication is likely?
If you’re talking only to dispassionate perfect rationalists, sure, they will make the proper analogy – but dispassionate perfect rationalists don’t need analogies, you can just use bland alphanumeric variables.
Everybody else is likely to be at least someone misled by the approximately aleph-null prior cases in which the analogy was between cockroaches and some group of nonwhite people, and may instinctively pattern-match to “which group of nonwhites is this guy trying to say are like cockroaches?”, and then right there at the start of the next paragraph is a reference to nonwhite people in a negative context.
Scott is right; there is no way in which using that analogy does anything but reduce clarity of thought and communication, compared to saying the same thing without the cockroach analogy. And I am surprised that anyone fluent in colloquial English doesn’t know this.
It’s a metaphor not an analogy. And while in retrospect it seems to have distracted people I’m mostly disturbed by how shallowly it was read. There’s a difference between being alert for dog-whistles and a complete lack of reading comprehension.
Anyway, to help get us back on track, an actual analogy with no distracting baggage that I’m aware of:
“This is like when you go down to the local used car dealership, pick out a car, and as soon as you drive it off the lot you hear a loud bang. You walk back to the dealer who shrugs and says that cars make noises sometimes but really it’s no big deal just relax. That may be, but even an idiot could tell you that for every problem you see there are ten you won’t and that the dealer has every reason to lie about it.”
I’m like 90% sure that the New Years gropes were a Daesh operation. Refugees probably don’t have a whole lot of incentive to engage in that sort of behavior, but how about a group whose stated intention is to make the western world hate all Muslims so that Muslims will have no choice but to join Daesh in their war against the west? Sounds like they have a strong incentive to pull shit like this. In fact, it seems like they’ve explicitly stated their intention to pull shit like this.
So before you wax philosophical about all the benefits of xenophobia, consider what other kinds of agendas might benefit from xenophobia.
Maybe when people say “you’re helping the far right”, they really mean the Muslim far right.
He’s comparing rapes to cockroaches. He’s not even comparing rapists to cockroaches, which it’s hard to argue wouldn’t be justified.
You’re more concerned with misreading his comment to suggest that he might have said something mean about migrants than with the fact that multiple European governments have covered up mass rape by migrants.
You can’t seriously believe that. Maybe you’re moral enough that you wouldn’t grope anyone even if you thought you could get away with it or if you were impulsive enough not to care, but there’s no way you’re stupid enough to believe that everyone on the planet is.
Refugees probably don’t have a whole lot of incentive to engage in that sort of behavior
Refugees as a whole don’t, but neither do they have a whole lot of incentive not to engage in it, especially since it is apparently without consequence. And certain elements of the refugees have more than one incentive to engage in it, and as mentioned, very little incentive not to. This is sort of a tawdry, disgusting version of that argument about “if you can do something that saves Hitler .0001 utils and harms nobody else, rationally you should do it if you believe in maximizing utility.”
@nydwracu: That isn’t what I did; I was responding to Alphaceph, and made the mistake of doing so with a simplistic scenario without addressing that the thing here didn’t actually match it. Responding was probably a mistake, and I apologise to Dr Dealgood for that.
I’m not sure what there is to say about the core points. It seems accurate that Cologne’s police did try to sweep their failure to act quickly under the rug quietly until external parties criticised them harshly for it and politicians got involved.
You could have a more detailed discussion, of the various factors leading to that, or elaboration on when they should have said what to who, but I don’t have detailed enough knowledge about what happened and what is normal or sensible process for police/news interactions, and who the individuals were and incentives on them, so I don’t have much to contribute there, and it isn’t necessary to simply make the high-level point.
I don’t really have anything to add to the thing about whether it counts as “mass rape” or not; I don’t know enough detail about what happened, Dr Dealgood’s perspective is reasonable enough, and I feel that going out of my way to try to find evidence for downplaying it would be biased; it isn’t something one would jump into conversations about other crimes of this sort to do.
And the basic point that this is a sign of a bigger underlying problem and probably more bad things than can be directly seen (underreporting, etc) seems pretty straightforwardly true, and would be the main reason for treating these cultural issues as a notable concern.
The only really contentious part is the claim about “European governments” in general, which is difficult to evaluate; it’s like generalising over “American states” but worse because they’re way more variable in governance and politics and much more independent.
The main issue with it is that I think they’re overestimating the size of the failed regional police/authorities relative to the size of their country’s government; these regions are maybe 1-10 times the size of a US county. There’s lots of them per country. You’ll get a lot of evil governments if you take every government which has had a regional authority hide from its problems until it fails drastically and put them on a list.
It’s a bit like saying that the US government is responsible for covering up Flint and therefore can’t be trusted- not exactly wrong to raise concern, per se, because the cultural and structural issues which led to Flint are probably latent elsewhere, but perhaps overly broad in its implication that the whole government was complicit in the corruption.
What interests me in this subthread is the question of whether you lose anything important by being unwilling to converse with people who deliberately violate norms of niceness. I think the answer is that you probably do, for two quite different reasons:
1. There are lots of not nice people in the world, and ignoring them makes it less likely you will understand them and so be able to make sense of their behavior. One of the benefits I get from arguing climate issues on Facebook is a corrective to my wildly optimistic view of the rationality and concern with reason of my fellow humans, based on the very nonrandom bubbles I have spent most of my life in. Reading trade chat in WoW provides a similar benefit.
2. There may be true things which existing norms of niceness make it difficult to say, in which case being willing to interact with people who don’t respect such norms, perhaps even enjoy violating them, make it more likely you will hear such things. For an old example, consider times and places where nice people didn’t talk about sex.
If I understand you correctly, you are claiming that:
-I cannot seriously believe that Daesh operatives have more incentive to commit sexual assaults given their stated intention of whipping up anti-Muslim sentiment than do honest-to-god refugees whose only incentive is maybe a fleeting sexual thrill, but who also have the disincentive of finding themselves in a new home where everyone is suspicious of them.
Typical mind fallacy or the like has nothing to do with it. The claim you’re making is absurd. Even granting the fact that many of the refugees may have poor impulse control, it remains true that poor impulse control is less of an incentive to commit sexual assault than is the desire to whip up anti-Muslim sentiment.
Even if the refugees have some small incentive (and I never said they didn’t) that says nothing about whether Daesh operatives have more incentive. Add in the fact that there were multiple sexual assaults in multiple cities all on the same night and it starts to sound a whole lot like a coordinated operation than the random action of refugees with poor impulse control even ignoring our relative assessment of their likely morality.
You’re far too quick to poo-poo my conclusion, and on the basis of invalid reasoning. That suggests to me that your conclusion that it’s probably the refugees is based on anti-refugee bias rather than a clear-thinking assessment of the likely causes the the incidents.
Most people are capable of reasoning: “I just moved into this neighborhood and nobody trusts me yet. I will have a better time over all if I make friends and establish a support system. That will be more difficult if I engage in illegal, immoral, or otherwise undesirable behavior than otherwise.” That is clearly a strong incentive not to commit sexual assault.
But even ignoring that, see above about Daesh having a stronger positive incentive than the refugees’ positive incentive (again, granting that the refugees’ incentive is at all positive, which I doubt it is).
A) No, my argument is nothing like that.
B) This is the kind of bullshit moralistic argument that this subthread is arguing against: “You can’t make an argument like that because it offends me!”
I tend to think you are also basing your opposition to my suggestion on anti-refugee bias, because as with my other interlocutor, your rejection of the proposal seems knee-jerk rather than based on valid reasons.
Edit: Another important point for you both to consider: the population of refugees is tiny compared to the population of white people in these countries, and the rate of sexual assault committed by white people is surely non-zero. How do we establish the base rates for sexual assault by whites vs. by brown furr-ner types? What sort of deviation from that base rate would we need to see before we should start suspecting a coordinated operation instead of random uncoordinated actions by a population strangely lacking in impulse control?
The portion of the refugees who have incentives to do these things (there are more than one) are certainly capable of following the logical chain of thought you describe. However, if you get to accuse me of bias, I get to accuse you of typical-mind fallacy: you are assuming they 1) agree with the logic, and 2) care. With respect to that portion to which I refer, I see no reason whatsoever to assume that either of those things is true.
I apologize, however, for implying that you were making the argument to maximizing Hitlertility. I did not mean to imply that (though I see in hindsight how it could be taken that way) and I don’t believe that you were. That is entirely due to poor communication on my part, and I am sorry.
What I meant by that was that since the portion of refugees to whom I refer have very weak if any disincentive to refrain from such acts, it doesn’t take a lot of incentive for it to become rational (from their point of view) to commit such acts.
As far as the base rates of comparative criminality of refugees, it’s probably too soon to have any trustworthy data, I readily admit. However, comparative rates of culturally similar immigrants have been studied, and the results usually don’t support the “they’re not more criminal than natives, you all are just watching them closer” hypothesis.
Even if that were true, if the consequences for equivalent criminal behavior were less for refugee criminals than for native criminals, that would be very, very bad, for reasons I hope I don’t have to explain. Again, it’s too soon to have any good data on that, but the comparison to culturally similar immigrants doesn’t fill me with hope on that score either.
No. Go back and read what I actually wrote.
I said that the chain of reasoning I gave supplies some negative incentive, but…
…and here’s the part you missed even though I made it clear as day…
…even if we ignore that negative incentive— even if we assume there is some small positive incentive from the sexual thrill of it! — Daesh still has more positive incentive than do the refugees.
And I specifically pointed out to nydwracu that the typical mind fallacy is irrelevant to my argument because, again, even if we assume the trivial positive incentive posed by the sexual thrill of committing sexual assault, Daesh still has a greater incentive to do this than do the refugees.
And that’s one of at least three arguments that I made for my position! Telling that you ignore the other two!
I accused you of bias so that you would do a better job of making an evidence-based argument to support your position instead of relying on “brown people make bad touch” noises to make your argument for you. If you don’t want to talk about base rates of sexual assault vs. population sizes and actual frequency of occurrence, then you’re only cementing my initial impressions of bias.
Edit: I appreciate the apology w/r/t the “tawdry, disgusting” bit, and apologize in turn about the fairly incendiary tone of the current comment.
“This person has more incentive to do something than that person” is certainly relevant, but hardly dispositive, when trying to figure out who actually did it. I understood your point: I just don’t find it all that compelling. I’ll see your Greater Incentive and raise you an Ockham’s Razor. 🙂
Your gracious apology is accepted, thank you.
“X has more of an incentive than Y to do Z” doesn’t mean that when you see Z, you should assume it is caused by X. You need to consider base rates.
You also need to consider mechanisms. The ability to organize a criminal conspiracy of this scale, without leaving smoking-gun proof when the police start looking into it after the fact, is rare to nonexistent in the real world. The IS’s motive for doing this pretty much evaporates if they get caught, and if they have a finite capability for organizing large criminal conspiracies without getting caught, they have far better things to do with it than organize a mass groping.
That razor tends to cut either way depending on one’s pre-existing biases. I’d tend to say Ockam’s Razor favors coordinated Daesh operation.
Population of Germany: 80 million
number of refugees in Germany: about 60,000, let’s exaggerate to 0.1 million
base rate of rape in Germany: 10/100,000. Let’s assume sexual assault is 10 times more common (100/100,000)
So we expect native Germans to have committed something like 80,000 sexual assaults over the course of the year. Let’s assume 1% happen on New Year’s Eve: 800 sexual assaults.
Let’s assume the Syrian refugees are 1000 times more likely to commit sexual assault than the average German. Then we have .1 million * 100,000/100,000 *1% (assuming again that an outsize proportion of sexual assaults occur on New Year’s Eve) = 1000.
So we need the base rate of sexual assault among the refugees to be incredibly high to break through the noise of background sexual assaults. Is 1000 a reasonable factor for this? Maybe, but I’d like to see more evidence for that before I believe it.
Or Daesh could pay, say 100 people to commit 10 sexual assaults each and very easily get the same number for a relatively small amount of money.
Can Daesh mobilize that sort of thing in Europe? Well, they mobilized a terrorist attack in Paris. This sort of thing exposes people to much less risk and probably requires a lot less organization to pull off (no smuggling weapons, etc.).
I’m curious what kind of evidence it would take to convince you that the sexual assault and rape issues in Germany are the result of calculated actions by Daesh instead of a general trend towards evil among brown people overall.
Yes, exactly! Talking about incentives and motives probabilifies my position, but we'd need to talk about base rates for a real evidence-based argument. I showed above that the base rate for sexual assault among the refugees needs to be about 1000 times the German base rate to get you a decent signal to noise ratio. I’m all ears on whether this is a reasonable estimate or for a rebuttal to this argument
How so? The only “criminal” part is the actual groping, which as others are pointing out is relatively low-risk. Organizing it is essentially zero-risk — you could organize something like this on Facebook! “Hey, go feel up some white women on New Years’, get me some video and I’ll send you $100 via paypal.” Oooh, what criminal masterminds we are dealing with!
Their motive doesn’t evaporate if they get caught, because frankly even if they got caught, there would be plenty of people who make specious arguments about why it must be the general tendency of brown people towards evil rather than geopolitically-inspired actions by one particular group of especially evil brown people.
It still achieves their aim of making white people more distrustful of Muslims. It’s not like a bank robbery where they fail unless they get every detail right. Pissing people off is not that hard to do.
The mistake in your calculations is the assumption that the assaults were independent. I don’t think anyone is arguing that. After the first attack, the chance for nearby people to join in is much higher than the base rate. There doesn’t need to be a general “trend towards evil among brown people”. All that is required is a small group of migrants starting the behaviour, and all the potential rapists will join in. This is much less unlikely.
Evidence that would convince me that the attacks were coordinated by IS would be any one of the >30 suspects being investigated said they were. What would convince you that they weren’t?
Or their metaphorical equivalent.
I can’t help but notice that your math, while arithmetically sound, makes a lot of assumptions that I don’t know are justified. As just one example, you are assuming more than one normal distribution without any justification for assigning a normal distribution to that parameter. Again, I understand it, and I don’t find it illogical, I just don’t find it convincing.
At a meta level, I can’t help but notice (and you are certainly far from the only or worst offender, nor your “side” likewise) that when someone complains about a particular event, their opponents are usually quick to point out that statistically, their concerns are foolish and could only be motivated by bias. Then when someone points out a troubling statistic, we are told that we should not punish individuals for the acts of a faceless population by using statistics to justify policy or explain outcomes.
Essentially, in this model, nothing is ever anybody’s fault, and any attempt to predict future results by past performance, at any level, is a fool’s game. That may actually be true, but in that case I don’t see why we bother to try to be consistent in the first place. Since neither individual acts nor the tendencies of populations are in any way helpful in predicting behavior, we might as well try just randomly punishing and/or rewarding people and see how that goes.
Cologne’s police failed to act quickly and tried to sweep it under the rug. Rotherham’s police deliberately covered for a Pakistani rape gang because to do otherwise would have been racist. Stockholm’s police covered up sexual assaults by migrants in 2014 for fear of “playing into SD’s hands”. And these are only three of the cases we know of — obviously successful coverups won’t be known.
I saw a lot of people try to talk about Rotherham like it was an isolated incident, but it wasn’t — there were, at the time of this writing, nine other migrant rape gangs in Britain that were large enough to merit Wikipedia articles. The Rochdale migrant rape gang was also covered up.
How common or rare do you think it is for people in the parts of the world these migrants are drawn from to commit random sexual assault for the hell of it? Do you think it’s, say, a common enough phenomenon to have a well-recognized name and several prominent campaigns against it?
The “same night” is the night when everyone goes out and parties on the streets and so on. It’s not like it was a random day picked out of a hat.
Right, which is why ethnic gangs of newly-arrived immigrants have never been a thing in America.
Can Islamic State pay a hundred people in let’s say ten major cities in Europe without any of the one thousand people they’ve paid going to the media and reporting that they got paid? How could they pull it off without any sign that they pulled it off? How could they have a Facebook page without anyone not involved realizing that they had a Facebook page?
Rotherham’s police deliberately covered for a Pakistani rape gang because to do otherwise would have been racist.
This is, by strong implication if not by explicit claim, well beyond what can be supported by any evidence I know of.
Many individual policemen (and social workers) in Rotherham disregarded individual accusations of rape against Pakistanis because they were afraid of being accused of racism. And didn’t talk about it because they were afraid of being accused of dereliction of duty. This behavior unfortunately obscured the scope of both rape and concealment of rape in Rotherham. There is little or no evidence that any Rotherham policeman knew of a mass “rape gang” that he was supposed to cover for, or that the Rotherham police as an institution knew of the tendency of individual Rotherham policemen to cover for rapists.
Cologne, Rotherham, et al are all readily understood as the collective behavior of individuals responding to incentives. The eagerness with which this is being turned into paired conspiracy theories, one of Islamic terrorists plotting mass rape for some ill-defined reason and another of European police forces plotting to let them get away with it, both unsupported by evidence, is disturbing enough when I see it in the press.. I am genuinely surprised and disappointed to see it in such an otherwise rational community as this.
Most of that description is pretty accurate, aside the clarification already made by John Schilling, and the bit about a cover up in the Rochdale case; there was a single case there that had an inquiry into why it was handled poorly/slowly. Political correctness was mooted as a theory, especially on the right, but investigations concluded the lack of resources for that case were due to a target-based culture focused on acquisitive crime (e.g. theft). This is pretty plausible, because “target-based culture breaks non-targeted thing” is a pretty common explanation for poor performance in the UK. While I’m sure people who didn’t agree with the results of the inquiry alleged a cover up, there doesn’t seem to have actually been any evidence of one.
That isn’t to say you’re wrong about the crime itself; the crime wasn’t an isolated incident except insofar as it getting to scale up due to being unconstrained. There’s sizeable (4-5x was one figure I saw) differences in rate of committing sexual offences between ethnicities. Social clusters where such crimes are normalised, and anything that supports their formation, are something to think about and look at means of disrupting. This doesn’t conflict with principles against collective assignment of guilt or punishment- doing it right probably even requires them.
For clarity I think I do need to point out: Rotherham and basically all the other UK cases haven’t involved many immigrants, let alone migrants-as-in-the-recent-migrant-crisis (who have very little presence in the UK). The involved people were mostly born as an ethnic minority within the UK. This doesn’t, you know, make the failures in Rotherham any better, but is a pretty important thing in other respects.
Phrases like “Pakistani” or “Asian” (which mostly means South Asian in the UK) in British media can be misleading; they refer to ethnic background and don’t say either way on whether they were born in the country or not. British Pakistanis are a sizeable population of apparently a bit over a million.
There’s very little presence of the recent migrant crisis, because France doesn’t let people smugglers operate on their coast, England has a moat, and so far UK politicians have refused to agree to accept transfer of any in Europe. The last bit may change, but is unlikely to involve large numbers; the Conservatives know their base is against it and are ideologically opposed themselves anyway. There’s people trying to get through via the tunnel or trucks, and causing trouble in Calais, but we’re talking a total of thousands trying, and the number succeeding being much smaller than that, rather than the tens or hundreds of thousands of arrivals in other EU countries.
Do members of secret intelligence organizations who are arrested for crimes committed in the line of duty usually explain to police that they committed those crimes as part of a subversive plot? Is a full confession the only thing that would convince you? Would you have held investigations of communist subversives to the same standard?
The matter is complicated by the fact that “refugees” and “IS operatives” are not mutually exclusive classes. Presumably, IS operatives can fill out the same application for asylum that any other refugee could.
My opinion is based on a criminal investigation framework: means, motive, and opportunity. It seems to me IS has all three in this case. However, a plausible explanation for why apparently coordinated attacks (coordinated because they happened across large geographic areas in a short period of time) weren’t actually coordinated would undermine the “means” and “opportunity” branches of that analysis and push me away from the conclusion that it was coordinated by IS.
I think it’s entirely plausible that an operation like this could be coordinated with no such evidence. Funds can be transmitted relatively anonymously through the internet, assuming the perpetrators weren’t already loyal enough not to need pay for a relatively low-risk and easy operation.
But at this point, I think your invocation of Ockham’s Razor indicates that you’ve chosen to go with your gut rather than seriously consider the notion that your opinion of Muslims is being manipulated to serve the purposes of IS.
Since this is the second time you’ve tried to typify my argument as something that it is nothing like, and since you’ve already essentially conceded that your opinion is based on your pre-existing biases w/r/t Syrian refugees, this strikes me as trying to dismiss my analysis so that you don’t have to seriously consider the idea that you’re carrying water for IS by believing exactly what they want you to. But please consider it anyway.
This shows that your previous paragraph is irrelevant nonsense and not an apt description of my position; I explicitly blame this on IS. It is their fault.
I concede that it is rather common. Now, do we conclude that all such migrants are rape-hungry monsters, or do we suppose that perhaps some small subset of them is responsible for the vast majority of sexual assaults? (Certainly, evidence on sexual assault in the US suggests it is the work of a small minority…#notallmen.) And if such assaults are committed by a small subset, do you think it is more likely that this subset is running away from IS or that they might actually be sympathetic to IS?
If it is a small subset, do we hold all refugees responsible for the actions of a small subset? How do you feel about that as a general principle of jurisprudence?
Sure, and I happily acknowledged that in my analysis, supposing that three times the base rate of sexual assaults occurs on New Year’s Eve. But I would expect similar events to happen every Friday and Saturday night, though perhaps in somewhat smaller numbers, if this was not a coordinated event. Do we see even 1/3rd of these numbers of assaults, robberies, carjackings, etc. on “party nights” besides New Year’s? How much of a deviation should we expect New Year’s to be?
On my understanding of ethnic gangs, they are typically composed of small subsets of the total immigrant population, and it’s often their own ethnic groups that compose their primary group of victims. Which is entirely consistent with what I’ve suggested so far: that a small subset of the Syrian refugees are IS agents or IS sympathizers willing to do favors or paid work for IS in Europe.
Similarly to how the Soviet Union could have thousands of operatives in the US and UK, including in the very highest ranks of their secret intelligence services. Or how all of Nazi Germany’s operatives in the UK could be turned by the UK secret intelligence services without Nazi Germany learning.
Edit: Also, since there are apparently less than 40 suspects, are we so sure this is thousands of people? Maybe more like dozens?
Facebook pages and groups can be made private, and there are thousands of private Facebook pages and groups — no one is paying anyone to monitor these things in real time. But I said they could have organized this on Facebook, not that they did. It seems likely to me that if I’m right about this they might have taken a few extra precautions.
“Would you have held investigations of communist subversives to the same standard?”
Whittaker Chambers was a communist operative who did make a full confession, and played a central role in the Hiss case.
It’s not even clear whether Scott means opinion polls, newspaper editorials, media spin/angle, celebrity opinion or all of these.
And then he attaches 60% to it – very close to the 50/50 ignorance split anyway. That 10% away from 50/50 is entirely swamped by the uncertainty in the definition of “mainstream opinion”.
It would be a lot more useful if Scott gave probabilities for easier to define/easier to falsify events, such as Merkel resigning, passport controls at a majority of European borders (where we carefully definite what counts as a majority), countries changing the number of immigrants they take in, etc.
You’ll know it when you’ll see it. For example, mainstream opinion is largely sympathetic to feminism, despite the majority of the public not identifying with it.
If there’s no solid criterion for judging it, then it’s not really a prediction, it’s just mood affiliation dressed up to look like prediction.
define “feminism” as used in that sentence.
Actually, it helps his argument if he doesn’t define feminism.
Most people get positive halo effect feelings from things labeled ‘feminist’. It doesn’t matter if the label has a meaning or not.
You can disagree with that claim, but the claim has nothing to do with how feminism is defined, just that most people define feminism in a way that aligns with their goals sufficiently to support it.
The vast majority of these are along the lines of “things will stay the same.” Personally, I don’t think that’s very impressive to predict.
But it’s almost always the correct prediction.
The interesting thing here isn’t the predictions but the confidence level. It’s almost always correct to predict that things will stay the same, but it’s important to have the correct level of confidence. “No civil war in the US” at 99.99% confidence is REALLY different from with 51% confidence.
It might be interesting to phrase your predictions as “[non-status-quo X] will obtain” and use the full 0-100% scale. Some prediction markets see more bias at one end than at the other this way.
(This also solves the 50% problem.)
I don’t understand the point of your low-confidence/low-information predictions. You claim your criterion is the accuracy of your confidence level, but it’s impossible to distinguish “my astounding insight has perfectly estimated the probability of each of this set of events at 50%” from, “I haven’t a freaking clue about any of these–could as easily be right as wrong, for all I know”. In fact, you can trivially jack up the accuracy of your 50% confidence level just by stuffing the list with a lot of completely random predictions that you know absolutely nothing about, and trusting the law of large numbers to bail you out. So why bother with any of them?
Way too optimistic on ISIS (still gonna hold Raqqa by the end of the year; thus is the will of Obama); too pessimistic on Assad and Trump in the general election. Romney sounded like a robot, he was totally unelectable in 2012; Trump sounds like someone who can inspire hope in millions of Midwestern voters. And, yes, his pivot will likely surprise many. And gas is $3 per gallon in CA. Wouldn’t want to live there; it’s too crowded.
Probably too optimistic on terror attacks; the IS warned of an attack in London recently.
Fights over “tiny stupid islands” should probably be marked subjective. Is Japan tiny and stupid?
Japan isn’t an island.
Did it become a continent when nobody was looking?
No, it became 4 islands that combine to form a giant robot.
I think he’s talking about the Diaoyu tai/Senkaku Islands.
The reason why, in ordinary writing, one hangs adjectives on a noun is to limit the reference to a subset of the things referred to by the noun.
It’s pretty clearly stuff like Dokdo/Takeshima, Senkaku/Diaoyu, and other tiny, uninhabited or nearly so, and disputed islands.
The confluence of Japan, US, South Korea and China as actors excludes the South China Sea disputed area, which would include China, Philippines, Vietnam, Malaysia.
FWIW, Scott, I’d coauthor a paper with you in a hot minute. Feel free to drop me an email.
My first reaction (and I suspect many others’ too) to that idea is, roughly: *SQUEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE*
I am entirely unashamed of this.
I’m curious about your terrorism predictions, and why you picked 100 casualties for your terrorist attack predictions.
I’m wonder on what you base your confidence that no such attack will occur in any first world nation, which would include Europe. To a lot of people, a major attack almost seems inevitable, especially with how horribly Germany has botched migration into Europe. Another Paris doesn’t seem very fair fetched.
I also wonder if your threshold of 100 people was a way to hedge your bets. You think there obviously will be terrorist attacks, but many smaller ones with casualty rates in the dozens instead of hundreds.
Yes, I’m talking about a large attack. Smaller ones are basically noise at this point. Oklahoma City, 9-11, Paris, Madrid, and various downed airliners hit the 100 person threshold, but it doesn’t happen very often. I suppose I’m claiming that recent events (including refugees and ISIS) don’t make this year too unusually dangerous.
Didn’t we have the discussion already on how giving 50% credence to either/or statements is useless for calibration? “50% I go to a wedding” is the exact same statement as “50% I don’t go to a wedding”, you have in fact made both predictions and will always get 1/2 on them.
Whether unemployment goes up or down, whether you attend any weddings or not, the “score” you give yourself at the end of the year on the 50% projections will depend solely on whether you phrased things in the positive or negative, not on any actual accuracy.
Tsk, tsk, the ministry of putting numbers on things doesn’t approve.
More importantly, 50% predictions mean “I have no idea what is going to happen, I have no information to distinguish these two things from a coin flip.” In Bayesian. So it’s more like saying “I conspicuously refuse to make a prediction on the following topic”
If he made a hundred 50% predictions on separate things, all with the same arbitrary-looking sort of phrasing choice as seen here, and they all turned out to be false, would that tell you anything about his calibration?
I had the same objection as you at first, but then I had another thought. All it really tells you is that he had really weird (bad) luck at choosing whether or not to put the predictions as negative or positive.
I don’t think that’s right.
50% predictions on binary outcomes can still be meaningfully judged in terms of calibration.
Take the fraction of 50-50 binary predictions that occurred*. Is it approximately 50%? If it is not then your calibration is off for 50%.
It’s true that ex-post it matters whether you input it as X is 50% vs. NOT X is 50% (in that one will be ‘correct’ and the other ‘incorrect’). But ex-ante at the time you are making the prediction you should be indifferent between the two equivalent predictions, so in expectation it doesn’t matter which you choose to record.
That’s not all it would tell me. Getting a hundred out of a hundred 50/50 predictions wrong in that way would make me think that when he wrote “I think there’s a 50% chance of this,” he was talking about something that he had strong reason to believe was unlikely.
One way to think about these probability calibration things is to think of “90% likelihood” and so on as arbitrary strings people assign to possibilities without much correspondence to their actual likelihood. The goal is, through practice, to train your ability to match “n% likelihood” with a belief that the thing is in fact n% likely.
If I have enough 50% predictions, the randomness of whether I say yes or no is just luck, and we see whether my calibration is okay or not.
To put it a different way – do you agree I can do meaningful calibrations on a probability of 50.01%? If so, why does it magically stop being meaningful when we lose 0.01%?
Yes, the particular phrasing you choose is exactly luck, and that’s what will be tested when you check the predictions, not your calibration.
Let an event you judge to be 50% likely be called E, and it’s “actual” probability be p(E), its converse !E, etc. Because you think that p(E) and p(!E) are roughly equivalent, your predictions are randomly distributed between the two, so your expected score will be:
.5*p(E) + .5*p(!E) = .5*(p(E)+p(!E)) = .5
This analysis does not depend on the values of p(E) and p(!E).
If you’re randomly picking whether to predict A or ~A, aren’t you always going to end up approaching 50% accuracy no matter what confidence you indicate? For example, if you say that something will occur with 90% probability, and you randomly decide whether that’s going to be A or ~A, you’re going to be right about 50% of the time. Similarly, if you say that something will occur with 50% probability, and you randomly choose whether it will be A or ~A, you’re going to be right about 50% of the time.
If I’m making an obvious logical error here, I blame it on the late night posting.
EDIT: I now realize that I’m saying pretty much the same thing as Montfort says in the post above mine.
I do not agree that you can do meaningful calibrations on a probability of 50.01%. Doing so would require something on the order of 10,000 predictions, which you did not make (otherwise, 50.01% is indistinguishable from 50%, so you can score perfect calibration by randomly flipping half your statements to their negation).
There is no number of predictions that lets you calibrate 50%; that doesn’t make sense.
It is obvious that a prediction of 50% is not confirmed by any result, and as you gradually get away from 50% in either direction, a result matching your prediction gradually confirms it more and more. I would bet there is mathematical literature about this. I’m not a mathematician or statistician and I don’t know exactly what the literature would say, but it’s got to be out there, and if Scott is going to make predictions, he really should try to track down the proper way to determine how well a result confirms a prediction.
Sitting here and trying to figure it out ourselves is, at best, going to rediscover an idea that almost certainly has been analyzed in detail already by people whose job it is to do such things.
It is our job to do such things, because math is fun!
One way to look at this is good old hypothesis testing. The null hypothesis is “I’m a completely cluless predictor who could’ve just as easily flipped a coin on each prediction”. You can measure good calibration if you have enough observation can distinguish the proportion you were looking for from 50%.
If you made N predictions with credence P in each one, then getting exactly P*N should be statistically different from 0.5*N. You can plug the following into Excel: ‘=BINOM.DIST(P*N,N,0.5,1)’, this gives you the cumulative binomial distribution. If it’s equal to less that 95% (or whatever your cutoff is), you can’t reject the hypothesis that you hit on 50% of your predictions and not P.
anon85 was off by a bit: for 50.1% you’d need around 680,000 predictions! Even at 60% you need around 60 to get there.
> do you agree I can do meaningful calibrations on a probability of 50.01%? If so, why does it magically stop being meaningful when we lose 0.01%?
Because you hit an asymptote at 50%. The number of predictions you’d need to make to reach statistical significance goes to infinity.
An analogy: Do you agree that the fraction 1/(x – 0.5) is well-defined for x=0.5001? If so, why does it magically stop being meaningful when we substract 0.0001 from x?
This is precisely the correct reply. I hope Scott happens across this comment and gives it a moment to percolate. It’s not a metaphor — the number of samples to reach an inference about calibration really does go to infinity at 50% asymptotically.
Interesting note: this is partially because True/False statements cover all outcomes in exactly two options — you could use conditionals to avoid this, like, 50% chance I go to no weddings, conditional I do go to one 50% chance it’s only one, condition I go to more than one, 50% chance I go to only 2 ad infinitum, which uses 50% probabilities for True/Falses, but chops up possibility space smaller so we can be objectively surprised when you go to 4 weddings.
This simply can’t be true.
If I hand you what I claim is a fair coin, you do not need to flip it an infinite number of times to believe or disbelieve me.
How so? Flip the coin any finite number of times and it might come up heads all the time. That’s what the law of large numbers is all about.
The appropriate comparison is, if you handed me 10,000 coins, where the sides all had unique labels (coin one has sides A and B, coin two has sides C and D, etc.) and I was only allowed to flip them once, I would not be able to tell if they were fair coins. Which is true. [[ The 50.01% case means the sides aren’t totally interchangeable from one coin to the next ]]
The difference between these scenarios isn’t necessarily obvious, and we can discuss that, assuming we are on the same page about if you give me 10,000 totally different uniquely labelled coins I can only flip once.
There’s a bit of confusion here, I should clarify what I meant.
Let’s say I claim I am an expert in politics, and try to convince you of that fact. I let you ask me 100 questions, and give you an 80% confident answer for each one. If I get about 80 of them right, you now know 2 things:
1) I actually have knowledge in politics (because you can’t fake this result by just flipping coins and guessing)
2) I am very well calibrated, since I know exactly how sure to be of my claims.
Now if instead I gave you 60% confident answers, the conclusion would be the same, but it would be weaker (i.e. it is still implausible that I just guessed and got lucky, but it’s less implausible). So to reach the same level of confidence in my abilities, you need to ask me more than 100 questions.
If we iterate this with smaller and smaller confidence, e.g. 55%, 50.01%, 50.00001%, it takes more and more questions to be sure I’m not just guessing, and finally at 50% there is no amount of questions that will convince you that I have any knowledge at all.
Surely Alex the raison d’etre of probability as an area of mathematics is to quantify the difference between “I flipped a coin once and it turned up heads, maybe it always does that” and “I flipped a coin a million times and it turned up heads every time, maybe it always does that”. I have no formal training but I’d love to hear from someone who does; in some sense in the latter case can’t we conclude with higher confidence that the coin is not fair?
It’s not that simple. cf. John Schilling below.
The deeper problem here is that if the stated purpose of making predictions at all is self-calibration or rather to prove well-calibratedness to the audience, then the person doing the predictions should not get to choose what predictions to make. Otherwise it is trivial to choose a set of predictions that maximizes observed calibratedness.
And, perhaps more importantly, this holds true, even if the list solely consists of high information value predictions.
Yes I’m aware of these issues, that’s why in my scenario the questions are not chosen by the answerer. I’m just explaining the asymptote thing.
I actually explain the problem you mention in a separate comment:
Right. If a predictor is cherrypicking their predictions, they can achieve any finite level of accuracy in either the predictions or the calibration of the predictions. This is independent of the values being predicted; others have noted how one can cherrypick a perfectly-calibrated set of e.g. 80% predictions almost as easily as for 50%.
If the predictions are not cherrypicked, even claims of 50% can have real predictive power and can be calibrated. And in this context, “here are some questions that seemed interesting to me off the top of my head” is if true as un-cherrypicked as “here is the official List of Predictive Power Calibration Test Questions that have just been removed from the sealed envelope by the accountants from Price-Waterhouse”. The latter is just more secure against fraud if you think that’s what’s actually going on.
I would not call it fraud. Rather I suspect that “questions that seemed interesting to me off the top of my head” is highly correlated with “questions on which I am well-calibrated” and therefore observed calibratedness on these questions will be a very very bad predictor of calibratedness on a random set of questions.
Why would I be interested in someone’s calibration on a truly random set of questions? I’d guess that just about anyone would have a predictive ability only slightly better than chance for questions drawn at random from the entire domain of human knowledge.
I’m interested in calibrating their accuracy on the predictions I am actually going to pay attention to. And unless they are being specifically hired to make some predictions, the predictions I am going to pay attention to are (a subset of) the predictions they chose to make for their own reasons. So that’s the set I want to see calibrated.
And part of what I want to calibrate is their ability to select questions within their field of competence.
We could get all the people who predict things to test/demonstrate calibration to also give probabilities for all of each other’s events. (Something like each throw in 50-100 questions which are non-personal and everyone gives a probability for all of them.) Still has the caveat of being biased towards areas of expertise on the rationalist blogosphere.
You can show that the likelihood of you hitting a particular outcome on your coin is somewhere between 50% and 50.x% with the value of x being dependent on the sample size of your flips.
But you can’t experimentally prove that a coin is perfectly fair. you can only prove that is is fair enough.
EDIT: someone explains it better below. First comment fail!
A claim that “half of the following events will occur” is exactly a claim that my predictions can be used in place of a fair coin.
The same techniques that you would use to estimate the fairness of a coin can be used to estimate this claim’s validity. Going further, any statement of the form “the following data will have this specific distribution” is testable in the same fashion.
(And let’s not waste time talking about the impossibility of achieving 100% certainty using probabilistic methods. Of course you can’t get that. And of course that doesn’t matter here. For every level of certainty below 100% there is a sample size that will let you reach it.)
We seem to have very different conceptions of what calibration actually means.
First, calibration is not predictive ability. On a random sample of questions one might not do significantly better than random guessing but that is not the point. To demonstrate calibration one has to be able to tell on a per question basis within the random sample on which questions they will actually do better than random guessing.
Second, assingning a 50% confidence to a prediction with a binary outcome is the same as admitting, that one knows so little of the field that one can literally do no better than flip a coin on that question. Conversely if one makes a lot of predictions of this class, one did not show ones ability to pick predictions within their field of competence. In fact confidence in ones own predictions is a perefect proxy of how close the question is to the core of their competence and 50% on a binary outcome means infinetely far away.
Third, maybe this explains the former two points, I find it very disturbing, how probability of the event and confidence in the prediction get intermixed throughout this discussion. Let me illustrate this by example:
If I were to predict the outcome of a coinflip and I wanted to express my opinion that it is a fair coin, I’d have to say that I am 100% confident that there is a 50% probability that the outcome is heads. Now, instead of flipping the coin, Alice gets to choose one side, and I want to express the idea that she has a strong preference for one side, but I do not know which side, I’d have to say that she will choose heads, but I am only 50% confident about that.
A yes I do get that in both scenarios I model P(heads)=0.5. However, this does not imply that the models are equivalent for the purpose of calibration. This is because, as I tried to explain in the first point, one has to measure predictive ability and calibration as distict and to that end one needs to know both estimates.
I was making the technical point that if redxaxder indeed handed me a coin claiming it was fair, there would be no finite amount of coin flips that could prove this to me with absolute certainty.
There are many ways in which this could go wrong:
1) [This is the one I alluded to in my original comment:] I do flip 10 times and get 10 heads. This is highly suspicious but not at all impossible. redxaxder claims incredibly “bad luck”. How do I choose if to believe him or not? Note how this is different from finding a coin in the street without any prior claims by a third party. Also note how, once we got to this point, an 11th flip of the coin can no longer save us.
2) Suppose I commited to flipping the coin exactly n times for large n and #heads/n = .5+epsilon for |epsilon|>0 (as it surely will be). On what grounds would I assume that epsilon, however small, is not in fact the coin’s bias? I have only redxaxder’s word to go on. Note that in real world scenarios, I would call “redxaxders word” a “theory” and be convinced. But it would be the combination of theory and experiment that convinced me the coin was fair. The coin itself has no way to prove this to me.
Absolute certainty is a type of brain damage. A state of mind in which no amount of future evidence can budge your confidence. Let’s not hold this up as our ideal.
A state of mind in which no amount of future evidence can budge your confidence.
redxaxder, I can be 100% confident I will not get married (to take one example). Now, it may be possible (as in “it’s not physically impossible, merely extremely improbable”) that I am kidnapped by someone, drugged or otherwise rendered suggestible and mentally pliant, and in that state we have our registry office ceremony and I end up legally married.
I’m putting confidence levels on that happening as 0.001%, just to give it a fair shake (sure, it could happen – and pigs might fly).
I can put 100% confidence on:
No-one will ask me to marry them
Even if I get a proposal, I will not accept it
In however many years of life left to me (be they 20 or 30), I will not change my mind on marriage nor will an opportunity to marry be presented to me
Call that brain damage if you like, but I’m unbudgeable 🙂
There’s something else interesting going on with the calibration at 50%. Presuming that you have make the positive prediction the choice you’d prefer (you’d prefer to be able to go to that wedding in Ireland, but you don’t know if you’ll have the money or time, for example), then being significantly off of 50% correct in the 50% predictions tells you something about your bias towards optimism or pessimism.
Just use a Brier score for all your predictions.
If you only make 50% predictions it is literally impossible to say anything about your skill at prediction.
If I point to forty six-sided dice and say “those dice are cursed and have a 50% chance of rolling ‘1’, but the sixty dice over there are fair and have the usual 1 in 6 chance”, and indeed twenty of the forty cursed dice and ten of the rest roll ‘1’, it would intuitively seem that I have made a more powerful and more accurate prediction than if I had simply pointed to sixty dice and predicted a fair distribution. Are you really saying that it is mathematically impossible to calibrate or otherwise recognize this predictive power simply on account of it resting on accurate 50% predictions?
Or perhaps more relevantly, if I select twenty stable first-world nations and say “each of these has a 50% chance of a coup this year”, and there are ten coups, I am either a superpredictor or a master plotter.
Those calibration examples require comparison to different sets of the same predictions. For your coup example, it works, but only compared to other sets of predictions. If another person predicted chances of coups between 1 and 10%, then we could say that it was awesomely calibrated compared to that person’s prediction. And you’re assuming (correctly) that I’m predicting low coup chances.
Predictions about if Scott goes to a wedding don’t have this property.
I agree with Furslid: This only works because you are using repeated events in your example, making it clear how to combine them into a meaningful statistic. What Scott is doing is more the equivalent of flipping two coins repeatedly. He flips the first, and if it comes up tails, he predicts tails, and if it comes up heads, he predicts heads, and then he flips the other coin, and records whether the prediction was accurate. Then he uses the results from the second coin to try to determine whether it’s a fair coin, having no way of knowing whether the first coin is fair. Also, the second coin is really a large set of coins, each of which he flips only once, but he treats it as though it were a single coin.
@ Furslid and Seth: I used repeated, common predictions for simplicity, not because it is necessary for the argument. So here’s the long version:
50% that there is a successful coup in Germany this year
50% that the Dow opens at exactly 17,000.00 for every trading day in March
50% that a particular coin will come up heads in each of the next ten flips
50% that North Korea will invade and conquer Japan this year
50% that Vladimir Putin will confess to the Litvinenko assassination and resign
50% that a single terrorist attack will kill ten million people in 2016
50% that he will find twelve new girlfriends, including Jennifer Lawrence and Emma Watson.
50% that an alien spaceship will land on the White House lawn.
Nobody else makes any prediction regarding these events, except to say “that can’t be right” to Scott’s predictions.
Four of these eight things happen.
You two combine your mighty brains and predict,
50% that Hillary Clinton will be the next US president
50% that a particular coin toss will come up heads
50% that the Dow will be down no more than 20% at the end of the year
50% that ISIS will be diminished but not destroyed in 2016
50% that there will be a fatal airliner crash in the US or Europe
50% that oil will end the year lower than $50/bbl
50% that US unemployment will be lower at the end of the year than the beginning
50% that Scott will get no new girlfriends
Only one of these things comes to pass.
Are you really going to say that, in this scenario, it is mathematically impossible to assert that Scott is a more accurate predictor than you? Because if so, I’m kind of interested in the thing that is apparently called “not math” that enables me to recognize that I should be taking investment (and dating) advice from hypothetical-Scott and not from you two.
Behold my “well-calibrated predictions”:
50% that there is a successful coup in Germany this year
50% that the Dow opens at exactly 17,000.00 for every trading day in March
50% that a particular coin will not come up ten heads in a row on the next ten flips
50% that North Korea not will invade and conquer Japan this year
50% that Vladimir Putin will confess to the Litvinenko assassination and resign
50% that a single terrorist attack will kill ten million people in 2016
50% that Scott will not find twelve new girlfriends who include both Jennifer Lawrence and Emma Watson.
50% that an alien spaceship will not land on the White House lawn.
50% that Hillary Clinton will be the next US president
50% that a particular coin toss will not come up heads
50% that the Dow will be down no more than 20% at the end of the year
50% that ISIS will be either be undiminished or destroyed in 2016
50% that there will be a fatal airliner crash in the US or Europe
50% that oil will end the year higher than $50/bbl
50% that US unemployment will be lower at the end of the year than the beginning
50% that Scott will get at least one new girlfriend
Your example of amazing 50% predictions has the same issue as saying that 20 specific stable countries each have a 50% probability of a coup: it is abundantly clear that the predictions were not flipped randomly because the amazing direction was always the positive prediction.
I’m not accusing Scott of willfully cheating, but only of being haphazard with the direction of his predictions, such that his measured calibration will accordingly be haphazard on the 50% predictions. He would need a pre-decided, unambiguous method of deciding the direction of the predictions for it to be otherwise.
It’s actually even easier to game the system than that, isn’t it? You can do it with any confidence level.
75% chance that [essentially guaranteed event happens]
75% chance that [another essentially guaranteed event happens]
75% chance that [another essentially guaranteed event happens]
75% chance that [some impossible event happens]
Boom, 3/4 of the predictions come true. I’m accurately calibrated!
With a little math, you could make it a lot less obvious than this and still score very highly. Ultimately you have to subjectively assess whether the person is abusing the system like this.
50% is special in that you can cancel out the weight of the statement without knowing anything about it.
The predictor flips a coin to decide whether to turn each statement into a “will” or a “will not”. In the end, this ends up having the best possible calibration for the 50% confidence level (assuming the coin was fair).
With your system, Gavin, you still need to make some predictions.
50% that North Korea will invade and conquer Japan this year
50% that Vladimir Putin will confess to the Litvinenko assassination and resign
Ah come on, John, those are bad examples. Everyone can be pretty sure that the confidence in those things happening is lower than 50%, even a lot lower (whether you want to go to 5% or 0.005% is up to individual choice).
50% is “I’m not very sure about this, I haven’t enough information or the information I do have is so conflicting it’s not possible to extract a definite conclusion, but at the same time I don’t feel confident enough to raise/lower the value”.
50% is “Will this omega 3 fish oil supplement help my joint pains? Who knows, but it can’t hurt to take it”. If I was really convinced it would, I’d say “80% it’ll help”. Having read some articles on “it doesn’t really make that much of a difference*”, if I were sure it’s useless, I’d say “30% it will help”.
Since I’m not sure one way or the other, but I don’t think it’s absolutely useless, then 50% I might as well take it.
*Or take the much-recommended ‘eat heart-healthy fats, like olive oil’ and the touted benefits of the “Mediterranean Diet”; imagine my surprise when my diabetes clinic nurse told me not to use olive oil but switch to something like rapeseed oil instead, as olive oil is bad for your triglycerides. So who do I believe? Googling shows that olive oil is supposed to lower your triglyceride levels, the nurse told me it raises them.
That’s a 50% confidence one right there 🙂
If you select twenty stable first-world nations and say “each of these has a 50% chance of a coup this year” and there are ten coups, that is very impressive.
If you give twenty predictions, each at 50% confidence, ten of the form “X will experience a coup this year” and ten of the form “X will not experience a coup this year”, where each X is a stable first-world nation, and there are ten correct predictions, that is not at all impressive.
The reason that the first set of predictions is more impressive is that the structure of the predictions makes it clear that the direction of each prediction was not selected at random.
The claim under debate is, “giving 50% credence to either/or statements is useless for calibration”.
Not “…some particular sets of either/or statements”, but the class of such statements generally. Any set of 50% either/or statements which could demonstrate predictive power, disproves the claim. No finite set of predictively trivial or useless statements will suffice to prove it.
If you’d like to offer anything suggesting a general solution to the problem of when and how a set of 50% predictions can be effectively calibrated, that would be useful.
OK, you are correct. I personally only claim that “giving 50% credence to either/or statements is mostly useless for calibration, and in particular is useless in the way that Scott does it.”
It would be along the lines of the end of my other post, “He would need a pre-decided, unambiguous method of deciding the direction of the predictions for it to be otherwise.”
One example of such a method would be: only give 50% predictions where the prior probability is sufficiently far from 50% that a betting market assigned less than 50% probability to it, or, in the absence of such, that almost everyone can agree that it’s less than a 50% probability.
There are two separate issues here.
Predictions of the form “I predict 1 outcome of N possible at 1/N probability” are not much of a prediction (especially when the trial can not be repeated). If you point at 400 six sided dice, say that each one will roll a 1 with 1/6 confidence it tells you nothing to very little. It certainly doesn’t tell you that the dice are fair. Every single one of them could be completely biased, but the pile is fair. And it definitely mean that ypou are “good at predicting dice rolls”. When you roll a six sided die one time, and predict some face at 1/6th confidence, you are saying “I have no idea which of these faces is likely to come up”.
I think the issue with the coup scenario is that the answers are all of the form “will have a coup”. We know something about coups, and we know they are unlikely so it is a ludicrous prediction to make. We know that the “I have no idea what the outcome is likely to be” prediction is not a good one.
But if I randomly predicted coup/not-coup, each with 50% confidence, for every single nation, I would very likely look like I was perfectly calibrated.
Okay, after reading everyone’s post I understand the objection better.
But there are still things that I am, in fact, 50% confident on, and I’m not going to ignore those, so the numbers stay.
But these numbers do not reflect 50% probability. You are forcing your numbers into the categories labeled 50,60,70,80,90,95,99%. You could just as well drop the 50% category, or switch to 55,65,75,90,95,99. Forcing a 50% to 45 or 55 is no worse than, in the current system, forcing 55 to 50 or 60.
That’s fine Scott. But saying 50% on a question with only two outcomes is the same as saying “I have no idea”.
So what you are calibrating is whether or not you really had no idea. Which is much harder than it would first appear.
Stockbroker says, “Yoyodyne, Weyland-Yutani, ACME, General Products, CHOAM, OCP, LexCorp, Rearden Industries, Sirius Cybernetics, and Wayne Enterprises each have a 50% chance of increasing their split-adjusted market price in tenfold 2016 YTY”. Five of these ten companies do indeed show 900+% gains in 2016.
Stockbroker makes the same prediction, with a different list of corporations, every year. And every year, 4-6 are indeed tenbaggers.
Market says, “…and after ten years, he still doesn’t have any idea what he’s talking about!”
“50% chance of increasing their split-adjusted market price in tenfold 2016 YTY”
The issue is that “50% means I have no idea” applies only to predictions where the number of possible outcomes (N) truly is 2. The objection you might raise is that the question being posed is a true/false question so N is, by definition two, but this is an invalid objection.
A simple example would be dice rolls of a six-sided die. Predicting a 6 with 50% confidence is not saying “I have no idea”. The same thing applies to predicting stocks will go up 10-fold.
If you merely predicted “up” vs. “not up”, rather than assigning a magnitude to the change, this would be similar to N=2. I think that making many picks all in one direction (up) starts to look different though. I want to say it is because they are not truly independent trials. The more picks you make, the more inter-dependent they are because you are measuring more and more whether the market is up.
Historically, markets are up more than they are down, so “no idea whether the market is going to be up or down” is actually a bad pick.
The objection you might raise is that the question being posed is a true/false question so N is, by definition two, but this is an invalid objection.
Why is it invalid?
Every possible real prediction can be reduced to a two-outcome true/false formulation, or expanded to many possible outcomes. Even a coin flip’s “heads” and “tails” are composites of, e.g., “land on heads without bouncing”, “settle on heads after one bounce”, “…two bounces”, etc., and really of a nigh-infinite number of possible trajectories.
What is the mathematical rule by which I may determine when a particular binning of states into “true” and “false” based on an arbitrary divide is “valid” or “invalid” for predictive or calibrative purposes?
As I’ve pointed out elewhere in this thread, the prediction “50% chance of increasing their split-adjusted market price in tenfold 2016 YTY” (no confidence given) is a very different animal than “100% chance of increasing their split-adjusted market price in tenfold 2016 YTY” (I’m 50% confident).
The former is useless for calibration of confidence in ones owns statements.
If I’m 100% certain, and correct, that the company will decide between a certain-tenbagger marketing strategy and a certain-bankruptcy marketing strategy based on the CEO’s toss of a fair coin next week, and I don’t feel like explaining sources and methods to outsiders, which prediction do I make?
Yeah. I very much understand that two different world models will lead to the same prediction here. That precisely is my point.
Using your world model to make predictions you might want to measure two very different things:
a) how accurate is my model.
b) how accurate is my level confidence in my model.
And now since we seem to agree that two diffrent models with two different levels of confidence lead to the same prediction, I conclude that predictions of this type are utterly useless to measure either one.
This is not to say, that the combination, as measured here, is not interesting. A large part of my objection admittedly is semantics. Why would one insist to use terms like “confidence” and “calibration” for what Scott does here? IMO ithat is incredibly misleading.
Oh, I think I finally have it sorted out:
Lets again consider my earlier example:
We have a coin. Alice thinks that it is a fair coin. So Alice makes the statement P(heads) = 0.5; 100% confidence. Bob thinks it is a biased coin, but he does not know in which direction it is biased so he makes the statement P(heads) = 1; 50% confidence.
Now we do not longer allow statements of probability but enforce absolute statements along the lines of Scott’s list. We ask both participants to predict the next 100 coin tosses.
In the light of the new events both participants decide to predict heads for all trials, at 50% probability because that is the only permissible enconding of their model.
We then actually carry out the 100 trials and conclude that the coin is in fact fair. Since Alice and Bob made the same prediction, we have to look deeper and find, that Alice, in hindsight, was right in her absolute confidence in her right model, and Bob in hindsight, was right in his minimal confidence in his wrong model. So both, Alice and Bob are very well calibrated rationalists. Yay.
But then again, should Alice not get some extra credit for actually getting the model right? Well, maybe not since this is about calibration. But can we at least agree, that Bob’s 50% certainty did indeed encode the (albeit correct) observation that he had no idea what was going on, and not some predictive power on his side that in reality only Alice has displayed?
“Why is it invalid?
Every possible real prediction can be reduced to a two-outcome true/false formulation, or expanded to many possible outcomes.”
I don’t think you are being charitable here.
Surely you admit to the difference in rolling a six-sided die and flipping a coin, as far as discrete outcomes are concerned?
Yes, there are minimally probably (literal) edge cases that we do not consider. These are cases where the coin or the die comes to rest on no particular side.
But do you accept that the right model for flipping a coin or rolling a dies is different, and that the difference between those two things is material?
Even as stupid & irresponsible as the Republican Party is, Trump will NOT get the nomination; the (manipulating) powers-that-be will NEVER allow it.
If you really believe this you can make good money by betting on the other candidates, since the current markets have Trump at ~45% to win the nom.
Which prediction markets would you recommend for US citizens looking to make “good money”, and what is your definition of such?
As far as I can tell, an American who wants to bet on the nominations without leaving the house is limited to PredictIt and iPredict. (For the general elections, there is also the Iowa Electronic Markets.) It is not possible to make “good money” on these. But an American interested in a large bet might consider flying to Vegas.
Yup, PredictIt maxes out at 850 per bet. Note that you can’t use it while overseas (their proxy for “are you from the US”).
I’m pretty certain that betting on the election is illegal in Vegas too, unless you were suggesting a illegal bookie or something.
Yeah, you’re right about Vegas. Maybe Canada?
Note what? I’m pretty sure PredictIt went through a lot of hoops to allow Americans in America to use it. Or are you saying America and NZ only? That would be bizarre. I’m not entirely sure about iPredict.
Are Vegas casinos allowed to take political bets? If so, do they generally? Large ones?
At 45% (5:4 odds) you will get 2.15 times as much money as you bet in winnings, plus a refund of your bet.
Not quite. 5/4 is pretty close to 0.45 implied odds, but it’s read as “For every 4 you bet, you get 5 if you win as well as your stake returned, nothing otherwise”. So at odds of 5/4 you’d be getting 2.25 times as much as you bet back if you win, *after* adding in the refund of your bet.
Decimal odds make this a little simpler; odds of X are simply read as “you trade your 1 for X conditional on winning, 0 otherwise”, and you expect a profit if your assigned odds are above 1/X. You can convert between the two by dividing the left side of fractional odds by the right, then adding 1 to it (so 5/4 turns into 2.25).
Getting 2.15 times plus original return, would be decimal odds of 3.15, which would correspond to implied odds somewhere below 33%- which makes sense, because if you assigned anything above that you’d expect a profit.
(I feel that I should point out that I don’t actually *gamble*; I’ve just spent a while pulling money out of bookmakers offering signup bonuses or promotional bets by matching operations at them with corresponding operations at exchanges. There’s tools for it and it’s a fun way to spend a little while before your accounts get closed.)
I suspect that such people as loudly announce that Trump is certain to lose are likely not doing so because they are convinced it’s true, rather because they want to help to make it true. As such, they won’t put any money on it, lest they fail.
Or because they are generally law-abiding American citizens who don’t want to deal with the risk and hassle of entering into a criminal conspiracy for this purpose.
If Trump wins both Iowa and New Hampshire, what do you think the powers-that-be will do to prevent him winning the nomination? Do they throw in with Cruz and pressure everyone else into dropping out?
At a minimum, they pressure everybody but Cruz and Rubio to drop out and make sure Cruz and Rubio chose delegates who will vote for either Cruz or Rubio over Trump come the convention.
Trump may seem a stronger nominee candidate than he really is, because the rest of the field is still so divided. The Establishment has not settled on their own candidate, so Trump’s opposition is fragmented and weakened by spending resources fighting amongst themselves. When the field gets winnowed after a few primaries, and it’s down to Trump versus one or two others, the Establishment backing may be much more effective. They don’t have to pressure anyone to drop out who is still polling single digits after early primaries – fundraising will likely dry up (excepting pet candidates of billionaires, so that may render the preceding idea invalid).
Think of it this way: Trump at say 40% support versus 6 others with 10% support each, looks one way. Trump at 40% support versus another candidate at 40% support with the remaining 20% up for grabs, looks quite another way.
I don’t like Trump, but if I were a Republican establishment high muckety-muck, I think he would be one of my top choices after, perhaps, Rubio, at this point for the following reasons:
Trump seems to tap into some sort of populist fervor which McCain and Romney failed to tap into and which Jeb!, Kasich, and, to a lesser extent, Christie probably can’t hope to tap into. Trump has a greater chance of bringing in those disaffected voters whose staying home cost McCain and Romney.
BUT, unlike the other “non-establishment” candidates like Ted Cruz, Trump, when you really look at his positions, is not all that conservative. What’s more, he has no political experience and, while long on bluster, could probably be expected to mostly toe the line if he actually got into office. He would make a big show of being different, but would largely do as he’s told by the relevant “experts”
The Democrats do hate him, but no more than they hate Ted Cruz, who, at this point, seems to be the only other “non-establishment,” “works up the base” candidate in a position to win. Plus, the establishment hates Cruz for his proven track record of refusing to play ball, wait his turn, etc. Trump is more likely to win the general than Cruz and more likely to toe the line if he does win.
And I think you see this happening in reality: despite NR et al, the establishment has thus far not moved to strongly embrace Rubio or strongly demolish Trump as one might have predicted they would. Though many are not ready to come out and embrace Trump, I think they’re adopting a wait-and-see attitude. If Trump wins a bunch of early primary states and momentum and enthusiasm seem to be on his side, they might chose to jump on the train rather than fight a losing battle for a nominee who would, in any case, lose to Hillary anyway.
Why would you expect him to toe the line or do what he’s told by the relevant experts?
Because I think he cares more about personal aggrandizement/success than about the advancement of any particular ideological or policy agenda. In order to be perceived as a “successful” president, especially by his own standards, he will need to “get things done,” not just cause a big logjam as say, President Ted Cruz might (a horrific log jam might actually be better for the country imo, but that’s another question).
In order to “get stuff done,” like say, build a crazy wall on the Mexican border, he will have to “play ball” to some extent. He may be an expert negotiator, but he is not an expert political wrangler. Even assuming the best of intentions, he would have to surround himself by and listen to policy and political experts to make up for his own inexperience. Given his own non-ideological nature, I think he’s all the more likely to listen to whichever experts he chooses, precisely because he cares more about being perceived as a success than about the advancement of any particular policy agenda. If the experts tell him “do x and the people will love you,” I predict he’ll do x almost every time.
One might object that he has thus far run a very unorthodox campaign and does not seem to have listened to political experts, but I think running a campaign and being the president are two very different things. Trump knows how to be a reality tv star, but that doesn’t mean he knows how to get policy passed. More importantly, I don’t think he cares much which policies get passed so long as he is perceived as a success.
It is certainly possible I’m wrong about how malleable Trump would prove while in office, but I think the Republican establishment higher ups think, at least, that Trump would be easier to manipulate than Ted Cruz, and I’d tend to agree with them.
IMO not so much “do what he’s told” as “leave the actual governing to the relevant experts.”
Because modeling based on narcissism, there’s little reason for him to get involved with specific policies. Getting into technical details – or even following through on his own broader ideas and statements – is not an efficient means of capturing public attention. You’d expect him to mostly continue to concentrate on getting on TV, while leaving the governing to others. Especially since this allows him the chance to occasionally and very publicly fire the people actually doing the governing – replace the leadership of whole departments, maybe several times over! – which he’ll enjoy.
The most problematic area would be foreign policy and diplomacy, where “saying stuff to get on TV” and “implementing policy” are often not distinguishable. He might actually start wars via provocative press conferences.
“IMO not so much ‘do what he’s told’ as ‘leave the actual governing to the relevant experts.'”
That may be a better way to put it. 90%+ of the day-to-day governing tends to go on regardless of who’s president. The establishment has a very strong incentive to see that, regardless of who sits atop the mountain, the foundations of the mountain go undisturbed.
This is why I think many Republicans genuinely would have preferred a 2012 Obama victory to a Ron Paul victory: having Obama as president means they can go on complaining as usual without having to actually do anything different. Having a Ron Paul-type as president could threaten the very foundations of the political class’s continued existence in its current form.
The establishment would like to have one of their own (i. e. Jeb!) sitting atop the mountain, but if they can’t have that, they’d rather have someone who will do a lot of grandstanding but leave the practical side of things to them (Trump), or even a Democrat.
Right now it looks like a 90%+ chance that the nominee will be Rubio, Trump, or Cruz. The establishment would rather Rubio than Trump, but I think they prefer Trump to Cruz. At this point, even if they succeed in discrediting Trump, most of his votes will probably go to Cruz rather than Rubio. Therefore, if your preferred order is Rubio>Trump>Cruz, as I think it is for them, it is better to wait for Trump to destroy Cruz and then let Rubio edge out Trump at the national level. Best case scenario you get Rubio, or, if he proves nonviable, maybe even Jeb! or Christie; worst case scenario, you get Trump. At least you don’t get Cruz. I think some of the GOP would secretly rather deal with Hillary than Cruz.
There are a lot of changes a president can make, Supreme Court picks, executive orders, simply announcing that you will take a dive on various lawsuits.
Even stuff that looked minor in the past turned out to be a big deal (everything relating to Bill and Bush w.r.t. the banks).
That being said, its hard to imagine Trump shaking things up too much, he is pretty much centrist with the exception of immigration.
I’m not saying a president can’t be important or have a really big impact, but I don’t think Trump is likely to be such a president. And I think the establishment is maybe slowly realizing that too. Between Trump and Cruz, at least, they’d much rather deal with Trump.
I’m not so sure about that. The narcissists I’ve met want to feel powerful, important, and adored. (Speaking colloquially, as I assume you are; I’m not attempting to describe clinical narcissism here.) They aren’t satisfied with simply maximizing their time in the spotlight; they need to feel like they’re running the show, or at least more important to it than anyone else is. They’re willing to delegate things they don’t find interesting, but they don’t take advice well; it implies that they don’t already know everything about the subject.
If Trump is a narcissist in this sense — and I’m not sure he is — he won’t be satisfied with TV antics while leaving the governing to others. He will understand that Presidents are judged mainly on policy, and he’ll look for policy proposals to build his legacy on. Most probably they’ll be dramatic and not very well thought out, and most probably they’ll fail. In the best case, they’ll fail in a way that the circle of yes-men around him can plausibly cast as success: for example, a reform bill passes but gets so badly gutted by riders and amendments that it’s toothless. In the worst case, they’ll be an obvious failure, and Trump will feel like he needs to double down on them.
I think President Trump would be somewhat more likely than most of the other people running to do or attempt something really crazy or ill-advised. I think this is a big part of why the establishment initially hated him and tried to destroy him. I think at this point many might rather risk a small percentage of him doing something really stupid rather than see the disagreeable, and possibly also crazy Cruz win.
Trump’s propensity to do something crazy, ironically, may also be a big part of his appeal to the rank and file; and to the extent I’d rather see Trump win than Hillary, even to me. I hate the current US political status quo so much that I’d rather take a small chance of someone doing something really stupid and crazy (short of, you know, nuclear war) than see it continue in its current form.
And herein lies another irony of US politics today: the Democratic Party is absolutely the “conservative” party in the sense that it includes far more people who think the US government is mostly on the right track. Maybe we need higher taxes on the wealthy, more comprehensive health benefits, more tight regulation on Wall Street, or whatever, but very few Democrats seem interested in a fundamental realignment (even Sanders seems mostly to want to put our current government on steroids, though he probably wouldn’t see it that way). They want Cthulu to swim faster in the direction he’s been moving for decades, not make him do an about-face.
The rank-and-file, blue collar Republicans (but not, importantly, the Jeb Bush Republicans) however, are currently in a state of serious anger about the status quo, both of the state of politics in the US, and about their own party. And I think that anger is certainly justified insofar as the Republicans haven’t done much of anything they promised to do since Reagan.
For better or for worse, this inchoate anger has coalesced around Donald Trump. This is ironic, because, both culturally and in terms of his actual policy views, he doesn’t seem to have a lot in common with these blue collar red tribers (other than his strong stance on immigration). I wish it had coalesced instead around Rand Paul or someone better, but it is what it is.
What I’m saying is, for Trump voters, “who knows what he’ll do! He seems crazy!” isn’t a bug; it’s a feature. Many of them are angry enough that they’ll take crazy over the status quo, and, while I don’t like Trump, I can sympathize with that way of thinking.
I think these angry voters are going to be in for a big surprise if they get their wishes and Trump does something really crazy. The status quo is pretty damn good in the grand scheme of things. Even for people who are deeply unhappy because things used to be much better for them.
If come 2019 we have a recent failed military coup, double digit inflation and U3, 10 murders / year / 100,000, and doctors demanding cash upfront before treating people, I hope Trump voters are happy that they wiped the smug looks off the faces of those darned coastal elitists.
@onyomi — I’m not a big fan of the status quo myself, but what’s your mechanism for Trump doing any serious damage to it that lasts beyond his first term? From where I’m standing, it looks like the best he can do is convince the GOP establishment that it needs more bluster and less conciliation in its rhetoric, while the Dems are more convinced than ever that their enemies are evil and crazy. More likely, whatever he tries to do falls flat on its face, national prestige takes a hit, and both party establishments dig in further.
@brad — I’m having a hard time thinking of a plausible path to any of that. Whatever Trump wants to do, he needs to do it in the face of bipartisan skepticism and an entrenched bureaucracy. He could easily make himself or the country look really stupid, write a lot of checks with his mouth that his ass can’t cash; but he needs support to make nontrivial domestic policy changes or any serious foreign adventures, and he’s not going to get it easily even with a GOP Congress. (Which he won’t have for more than two years if he starts screwing up badly.) That implies either compromise or failure, and I think failure is the more likely.
So, gridlock. Well, we already know what federal gridlock looks like; we’ve just had four years of it.
Whatever Trump wants to do, he needs to do it in the face of bipartisan skepticism and an entrenched bureaucracy
This being Trump, it seems quite plausible he will respond to the inevitable failure by firing the bureaucrats. There’s no way that ends well.
I’m really just mostly engaging in conjecture about the psychology of the election. As for what exactly Trump would do if he got in, I honestly don’t have a very good sense at all. Which is kind of scary, yes, but for people who are currently pissed off it might be exciting and/or an opportunity to project their hopes and dreams on him. As Nornagest says, the most likely outcome of this election, regardless of who wins, is more gridlock. I just think that if Jeb Bush or Hillary Clinton have a 1% chance of doing something really weird and stupid, Donald Trump might have like, a 4% chance of that. Maybe he has a 20% chance of proposing something weird and stupid (or, arguably, 100%, given some of the things he’s already said), but maybe his chance of enacting it is 4%.
Correct me if I’m wrong, but I was under the impression that a sitting President can’t just unilaterally fire most of the executive branch. He can restructure it with Congressional support, but he’ll have a hard time getting that. Or he can reverse appointments, but appointed positions are a relatively thin crust on top of a much larger establishment.
The fact that we can confidently talk about what the president can and can’t do is a major asset for our country. I’d be very loathe to risk that even for a politician whose positions I mostly agreed with.
I agree, but what makes you think Trump has a chance of successfully subverting that? It’s been tried before, by Presidents much more popular than he has a realistic chance of being.
I was under the impression that a sitting President can’t just unilaterally fire most of the executive branch
It is not legal for a sitting President to unilaterally fire mid-level executive branch employees, correct. “Can’t” is a different thing from “not legal”, and therein lies the potential for real ugliness.
@Nornagest: Which presidents are you thinking of as having tried and failed to fire any great number of civil servants?
I should note that I also think that Trump has a higher percent chance than Hillary or Jeb! of surprising everyone and turning out to be a great president. Not super likely, but not impossible–lots of people thought Reagan wasn’t serious and was nothing but a movie star, recall (of course, many people aren’t happy about the legacy of Reagan, but he was obviously deemed a big success by his own party and constituency, if nothing else). I’m not saying Trump is the next Reagan, but he’s more likely to be than Jeb!
If Jeb! or Hillary wins you can be fairly sure they will do an okay job but that nothing very big will change barring exogenous shocks that have nothing to do with them. If Trump wins, there’s a slightly higher chance he’ll do something awful, but also a slightly higher chance he’ll do something really interesting and/or shift the political landscape in an ultimately positive way. And I’m sure Trump’s supporters rate that probability much higher than I do, of course.
Not trying and failing to fire lots of civil servants as such (though I wouldn’t be surprised to find that it had been tried), trying and failing to subvert the federal checks-and-balances structure. FDR’s court-packing bill is the one I was thinking of when I wrote that; there have been lots of similar episodes, some bolder than others. Obama’s recess-appointments loophole is the most recent one I can think of at the moment.
Offhand I can’t remember any that were blatantly illegal so much as creatively interpreted, but I don’t think that would be a point in Trump’s favor here.
Before running for POTUS, Ronald Reagan had proven his chops by being governor of California. He had a track record of governing, something Trump lacks. (And of course this is supposed to be a feature-not-a-bug according to his supporters.)
“He had a track record of governing,”
Yes, but that didn’t stop people from complaining at first that he wasn’t a “serious” candidate.
Don’t get me wrong: a Trump nomination would still be pretty unprecedented. I think he doesn’t know what the hell he’s doing and would probably be a bad president. But I wouldn’t bet a lot of money on his turning out to be a bad president. Maybe a little money.
Put another way, if I had to assign a probability to a theoretical Trump presidency being widely perceived as a failure, even by Republicans, within a few years of its ending (let’s say to the same or greater degree Bush Jr’s presidency is now largely perceived as a failure), I would only assign it maybe 60-70%.
It’s a somewhat ill defined concept when used retrospectively but I have no qualms about saying FDR was a dangerous President. Not just the Court packing plan either — his early actions on gold could easily have gone pear shaped (and I say that despite being as far from a gold bug as it is possible to be.)
FDR backed down on packing the Supreme Court, but he got what he wanted—the switch in time that saved nine. And one of his predecessors did in fact pack the court—add an additional Justice in order to make sure the court didn’t rule his activities to be illegal.
The clearest case I can think of of a president acting illegally and getting away with it would be Andrew Jackson’s relocation of the Cherokee.
FDR had victory in WWII as an end to retroactively justify any means he used to achieve it, and he conveniently died before he needed to come up with another justification for the imperial president-for-life bit. If the best we can say of Trump is that he might not be as bad as FDR, I think we’re going to need a World War in there somewhere. A real one, not just a few terrorists who stage out of half a dozen scattered countries around the world.
And Trump seems to be in good health, so are we expecting another presidential assassin for sometime in his second term?
Have I missed something? It seems like people are taking the idea that Trump is going to be some mad dictator a lot more seriously than I would have guessed.
I might vote for him in the general for a the humor value and as a “screw you” to the media-political classes but don’t think he’ll accomplish much. Carter showed what happens when you put a real Washington outsider in the White House: you don’t get reforms or coups, just a one-term lame duck presidency.
Carter didn’t have the culture of “executive orders are like laws made of dreams” that we have today to work with. If Trump decides to be the anti-Hope and go nuts with Executive Orders, it will be very interesting. Much like the filibuster, until very recently nobody wanted to rein in EO because they liked it when their guy used them. If Trump wins and starts slinging them around, they’ll either have to challenge the scope of EO legally, which goes nowhere good for either major party, or suck it up.
When the President does it, it’s not illegal. Seriously.
For something to be illegal, there needs to be a higher authority to enforce the rule. Who is going to stop him?
I have in my rss reader, this very day, a Washington Post article titled “The illegal implementation of Obamacare.” It’s from Volokh Conspiracy, so all very lawyerly and technical, and no doubt they make a decent legal case, but ultimately, who cares? The general public certainly won’t. There’s not even a court case involved.
I saw plenty of Bush’s actions called “illegal,” too, up to and including the Iraq War, which nevertheless occurred. I’m sure each president before that did many “illegal” things. Probably most of them could be argued (Was bombing Libya without congressional approval for multiple months illegal because it violates the War Powers Resolution? Or was it legal because Congress doesn’t have the authority to constrain the executive like they tried to with the WPR in the first place? Or maybe just because this bombing didn’t constitute war-war? How often do you even think of that whole thing, several years later?)
If Trump’s illegal actions consist of firing civil “servants,” then good on him. I’m sure somebody will find a method by which it is not technically illegal, and even if they don’t there’s really no controlling legal authority.
Trump trying to fire bureaucrats would actually be a very traditional illegal overstepping of Presidential authority. We have a rather famous court case that is pretty closely on point. I’m not an expert on this field of law, but I believe the Civil Service Reform Act of 1978 entitles Federal employees to sue for wrongful dismissal under certain circumstances. US District courts are empowered to order reinstatement and back pay. Ultimately, courts have writs of mandamus and the contempt power to enforce their will on recalcitrant officials, including POTUS. Obviously courts need the cooperation of other actors, like bailiffs, to enforce their will; but if we assume that Trumpism does not completely permeate the entire American system of government, it’s plausible that President Trump would encounter limits to his power.
This is exactly right. Dems are conservative relative to the New Deal consensus.
I very much disagree with this. “More tight regulation on wall street” is a huge turnaround on the current Washington consensus — it was Clinton who signed the bills deregulating the financial industry and it’s not as though Republican insiders disagreed with the idea (arguably Clinton did it because of the country’s rightward shift after Reagan).
If Sanders was actually capable of putting a wedge between the financial industry and the political establishment, then I would call it a huge structural change to our society and government. I would say it’s probably bigger than any of the immigration proposals that Trump has made that are actually legal in the first place.
I think the weird quasi-metaphysical ways you guys have devised of talking about politics may actually make your thinking on the subject less clear. They certainly make your communication on the subject less clear.
On the question of whether Trump could fire bureaucrats …
Suppose he doesn’t fire them in the usual sense of taking them off the payroll. He just cuts them out of the bureaucratic line of command—alters the organization of their part of the bureaucracy (presumably through someone he has appointed to be the head of it) in a way that leaves them with no power to affect any outcome.
Is that doable? Illegal?
As an alternative strategy, suppose his person at the top acts in a way that makes it obvious that keeping your job, if you aren’t on his team, is going to be a dead end–no promotions, the office without windows, general bad treatment. Think of it as analogous to how a university gets rid of a tenured professor they don’t like.
Are the legal constraints sufficient to prevent that sort of thing? Have past presidents done the equivalent when dealing with a bureaucracy largely populated by people hostile to their policies?
I thought this was a key component of the revolving door. New appointees come in every few years and announce radical new policy directions. Mid to top tier personnel either adapt the new way, or take an offer in industry.
Those that fail to adapt but can’t get an offer in industry are bad enough at their job they can be let go for poor performance, and a couple of principles holdouts can be sidelined into dead end positions. Firing the peons is not on anybody’s to-do list.
Edit: if you want to radically reduce staffing/eliminate whole departments to implement a libertarian utopia, take a page from de Blasio’s playbook: close the current facility and manadate the use of a new one that will never actually be built.
@David Friedman: That would be legal but expensive – basically the equivalent of a schoolteachers’ rubber room. No work, full pay, until they find something better or retire. It can be an effective means of neutralizing a recalcitrant or corrosive civil servant, but it isn’t generally effective at getting them to actually quit and it isn’t a viable strategy for winning a war with the civil service in general.
Among other things, it gives a bunch of disgruntled civil servants a bunch of fully-paid time to do pretty much whatever the hell they want so long as they can do it from inside an office. And the opposition is going to be looking for agitators, gadflies, campaign organizers…
(Yes, technically you can fire them for doing that sort of thing on government time, but it’s prohibitively difficult to prove and the civil servants know it – which is the source of the problem in the first place)
“If Sanders was actually capable…
I think the weird quasi-metaphysical ways you guys have devised of talking about politics…”
On the first point, I would agree that if Sanders were able to successfully put a wedge between the financial industry and the government (which, means, basically, disconnecting the performance of the stock market from politics and somehow disciplining or even abolishing the fed), it would be a very big deal. But I can’t think of any way he could do that without basically destroying and restructuring the whole US economy and political establishment. Any half measures would only result in adjustments and rearrangements of the current system.
Re. Cthulu, I thought “Cthulu always swims left” was well-known enough around here to be used as a shorthand, both for the New Deal economic consensus and for the Civil Rights-era cultural consensus. All I mean is that Dems want to keep pushing us down the trajectory the New Deal and Civil Rights movements set us on, while conservatives want to either roll back some or part of that and/or follow a different course entirely.
Just piping up to agree that the cthulhu stuff obscures more than it illuminates.
“Just piping up to agree that the cthulhu stuff obscures more than it illuminates.”
Oh, c’mon. Of all the obscure acronyms and LW jargon that gets thrown around here, with all the talk of “Moloch” and orthogonality, everyone’s going to complain about my reference to what I thought was, around here, a very well known bit of Moldbuggery? And I don’t think I was being needlessly obscure or cute: that phrase, though maybe not familiar to all possible readers, references a lot of history and assumptions and implications which I didn’t think I needed to rehearse here. If it’s more obscure because I didn’t insert a brief history of 20th c. political and social thought as seen through a reactionary lens, well, then I’m obscure.
Orthogonality is an actual word, being used in a sense pretty similar to its conventional usage.
It is not conventionally used outside mathematics.
If you google cthulhu swims left you get thousands of words of turgid prose. I didn’t read it the first time I googled it and I didn’t read it just now.
I got the gist of what you were saying but don’t know what it has to do with cthulthu or what exact nuance the reference was supposed to add.
Inasmuch as there was intended nuance I feel fairly certain you could express it more concisely, eloquently, and charitably than Moldbug did. You don’t want to, that’s certainly your prerogative.
FWIW a less wrong-ism is a good analogy. Google *that* and you’ll hit a post where ever other word is hyperlinked to another LW post, all of which assume you’ve read the Sequences of Prophet, Peace Be Upon Him. Ain’t nobody got time for that.
This is a long standing community and members have no obligation to clue in new guys to all the shared history, but at the same time given that it’s an open one you have to realize that you’re going to get some grumbling.
@David Friedman: Rubber rooms have already been mentioned. As for passing civil servants over for promotion etc., on the basis of political affiliation, all that stuff qualifies as “prohibited personnel action” under the CSRA. Which is not to say it never happens: the US Attorney scandal is one example. But while the US Attorney scandal didn’t result in criminal charges, but it’s hard to frame the scandal as a win for the Bush Administration. Had AG Gonzales not resigned, or had the purge been more wide-ranging, the legal proceedings would probably have progressed further. And that was only with US Attorneys, who are political appointees and have no legal protection against arbitrary firing! Had the Bush Administration actually tried to purge or punish protected civil servants, the story would likely have turned out much different.
I thought “Cthulu always swims left” was well-known enough around here to be used as a shorthand, both for the New Deal economic consensus and for the Civil Rights-era cultural consensus. All I mean is that Dems want to keep pushing us down the trajectory the New Deal and Civil Rights movements set us on
Is that all? I thought Moldbug or whoever was talking about a planet-wide Moral Fall which began long before “for loss of an eye, start a blood feud of seven generations” got regretably softened to “an eye for an eye”.
Then we’re in violent agreement on this point. 🙂
I guess my problem comes from the notion that “the New Deal economic consensus” and “the Civil Rights-era cultural consensus” together compose some monolithic quasi-metaphysical tendency for all societies everywhere. I don’t think that’s the case. I think both those phenomena are grab-bags of attitudes and policies, some of which are sensible and some of which are incredibly stupid.
I guess the problem is that Moldbug has substituted “anti-Whig history” for “Whig history”, but neither is a very good description of what actually happened/happens.
The Republican establishment has tried to demolish Trump, via the usual method of having party elders like Dick Cheney call him unfit to lead. It backfired on them, and they learned their lesson. You wouldn’t expect the rank-and-file GOP Congresscritter to go around denouncing anyone in the party. Reagan’s Eleventh Commandment and all that.
Also, I think people tend to underestimate the degree to which the GOP dislikes Trump due to substantive policy disagreements (e.g. Trump is not big on budget-cutting, wants to raise taxes on the rich, is unconvincingly pro-life, emphasizes immigration issues that many of the party’s elites would prefer not to deal with, etc.). It’s not just that he’s a maverick outsider challenging their dominance; he’s a very heterodox Republican. I think that’s where the Obama comparison falls down a little.
It’s important to clarify that in practice the electoral influence of the “establishment” under the modern system means, basically, the influence of the donor class. People tend to think of the establishment as politicians and bureaucrats but that’s not really where the electoral influence comes from. It comes from donors.
And you’re right – because of this mis-characterization, people underestimate the ideological commitment of the establishment. Big donors do tend to be ideologically motivated, even when politicians aren’t. And the Republican donor class does dislike Trump because they (rightly) perceive his lack of commitment to their principles.
On whether the establishment has tried to take down Trump – they haven’t. Remember that the establishment in this context is the donor class. An attempted takedown doesn’t look like party elders offering negative opinions to journalists. It looks like several hundred million dollars worth of targeted attack ads, and massively funded multistate campaign organization for a selected adversary. This hasn’t happened yet.
I should emphasize the YET. My perception is that the Republican establishment is still hoping to wait out the Trump wave. They don’t want to be perceived as having put their finger on the scale, lest they further alienate the already-alienated bloc that’s driving Trump’s popularity. If Trump loses with any sort of grievance, it’s likely to trigger a run as an independent and a subsequent general election disaster for the party.
And there’s the complicating factor that they still haven’t settled on an alternative candidate. Their preferred outcome is that Trump gets a “fair shot” but falls short in the primaries, with Trump voters returning to the party fold thereafter. I suspect the donor class is hoping current Trump supporters will help pick the eventual candidate, thus insuring they’re satisfied enough not to support an independent Trump candidacy.
The republican party doesn’t have nearly as many tools to fight it at their disposal as the democratic party and the democrats couldn’t keep Obama out in 08. They’ll fight it for sure, but that doesn’t mean they’ll win.
Hell, the Democrats have done a lot more than the Republicans to get rid of any candidate that wasn’t Hillary, and while not as unsuccessful as the GOP (Hillary does lead the polls), they still didn’t get it done.
I think that’s more “Clintons” than “Democrats,” as far as you can make them distinct.
As a nominal Democrat and plurality-D voter, I’m somewhere between distressed and angry with the way Clinton drained the Democratic well of all other mainstream candidates.
Sorry, I thought it was apparent that I was talking about “Democrat Elites”, that is, the same guys in the opposite team of the “Republican Elites” that were unable to stump the Trump.
And it is honestly kind of perplexing. I mean, as a foreigner, I kind of liked Webb, and I still fail to see exactly what’s wrong with O’malley. There was even meme potential there, with him having been played by CIA guy.
Dow Jones will not fall > 10% this year: 70%
The Dow was at 17,000 on the 4th at open, and closed in the 15,000’s today. Did you mean:
The DJ will not fall >10% from date of posting?
The DJ will not be down >10% from where it was at open on the 2nd January on the 31st December?
The DJ will not experience any closes >10% below that days open?
The DJ will not experience any closes >10% below where it was at open on the 2nd January?
Because the last one has already come true.
Howdy y’all. I’m a long time reader, first time commenter 😉
(Sidenote: I suppose I could mention, however, that I was actually at the recent Harvard SSC meetup. Specifically, I was the guy who — in the freeform discussion after your talk, Scott — mentioned the “Skills, not Pills!” motto of the Dialectical Behavioral Therapy partisans at McLean Hospital.)
Anyway, here’s my question:
In light of stories like this one today from the Washington Post (“As Zika virus spreads, El Salvador asks women not to get pregnant until 2018” https://www.washingtonpost.com/world/the_americas/as-zika-virus-spreads-el-salvador-asks-women-not-to-get-pregnant-until-2018/2016/01/22/1dc2dadc-c11f-11e5-98c8-7fab78677d51_story.html ), what do y’all think about the odds that…
1) … the world goes all _Children of Men_ this year?
2) … that there’s *another* pandemic illness in 2016 that is like Zika in that it impels one or more governments to issue public health warnings to, say, 10^7 or more people that match the really-freaky-scary scope of this Zika warning (“umm, out of prudence you might not want have children for the next year or two… nothing big… also, maybe get mosquito nets… they’re not just for malaria zones anymore!”)
3) … that there’s a pandemic in 2016 that kills more than 10^6 people?
In closing, for any of y’all who are now scared because you don’t think I’m just being paranoid about pandemics, well… I wholeheartedly apologize for being a downer. 🙁 For those of you who think I am paranoid, lemme defend myself as not really being that paranoid. My own probabilities for (1-3) are:
2) 5% in any given year to have a Zika-sized public health scare causing governments to strongly suggest, umm, major lifestyle changes (for lack of a better phrase). To have *two* such public health scares in a given year, I'd assume independence and say (5%)^2 = 0.25%.
3) For time-circumscribed mega-killing pandemics, I'd say the 1917-18 flu is really the only example, though please correct me if I'm being egregiously underinformed. Please note, however, that I'm only considering novel, time-circumscribed events, and thus I'm purposely not including the disease mortality burden from perennial diseases. On that note though, I suppose since AIDS was novel in the latter 20th century and quickly reached pandemic proportions before settling into being yet another piece of the world's perennial disease mortality burden, I could consider the 20th Century as having 2 such mega-pandemic events. Thus, in conclusion, I roughly estimate a 1-2% probability per year for a mega-killing pandemic. (Discuss amongst yourselves if increased economic interconnection of the world and more DIY-biotechnology combined with weakening social stigma against massive political violence against civilians means this annual risk will significantly increase in the near future.)
Third Pandemic plague used to kill a million people a year in early 1900s India alone. Totally unimpressive for Medieval plague, very impressive for the modern day. The infectious disease which caused the most human deaths in 2014 was tuberculosis, with 1.5 million deaths worldwide, followed by AIDS with 1.2 million (with overlap of 400,000 between these two). I think you’re going to have to upgrade that to 10^7.
Thank you, E. Harding. I very much take your point. As bad as it would be for a novel strain of a known infectious disease or a newly discovered infectious disease to kill 10^6 people this year, it’d still be in the same ballpark as the death totals from the perennial infectious diseases.
Indeed, I now see I’ve been underestimating worldwide flu-related deaths in a typical year by something like a factor of 5 as I now see that the WHO lists 250,000 – 500,000 flu-related deaths as a typical annual toll ( http://www.who.int/mediacentre/factsheets/fs211/en/ ).
Of course, talking about flu deaths in a typical year brings up arguments about Disability-Adjusted-Life-Years (“DALYs”) and such because — if I understand matters correctly — typical flu disproportionately kills the very young, the very old, and the respiratory-challenged, whereas pandemic flu strikes down everybody, including healthy people in the prime of life.
Alas, it’s getting past my bedtime, and besides, DALYs and such are not part of my expertise.*
[* Amusing Anecdote: I met my lovely wife through eHarmony,** and on our first date we bonded over discussion of DALYs, how low the threshold of $/DALY was for the WHO and similar bodies to view a public intervention as feasible, as well as lighter matters… you know, like pathologies with funny, rhyming names like “minky-kinky hair” and “tonic-clonic seizures”! She’s a psychiatrist who thought she might go into public health. I’m a mere physicist who once thought he might be a physician. Aaah, young nerdtastic love! 😉 ]
[ERRATUM 10:29pm – Oh my. I just Googled “minky kinky hair” and now see that I was mishearing the name all these years! It’s *Menkes* kinky hair syndrome.]
[** Obligatory Quasi-Disclaimer: Before y’all think eHarmony’s patented “29-dimensions of compatibility” algorithm (or whatever the heck they call it) is super-duper-needle-in-a-haystack-finding-amazing, I must note that I had much, much less compatible matches (indeed, some hilariously incompatible ones) with at least 10 other women before being matched with the woman I eventually married… though, come to think of it, that’s still pretty super-duper-needle-in-a-haystack-finding-amazing if our 10-matches-til-true-looooove story is at all representative of eHarmony’s algorithmic matchmaking prowess.]
How come we aren’t mentioning the Spanish flu here?
Oh, my bad.
Your odds on Clinton and Trump are high enough compared to the current market that you can make significant expected $ by betting on them.
It’s unfortunately illegal to do so for Americans, outside of a couple small restricted services that got an explicit waiver from the regulatory body. For example, PredictIt allows buying 850 dollar contracts on each bet. You can make a little change but definitely not significant money.
And it should be illegal. Allowing betting creates bad incentives.
Also, gambling is an attractive nuisance that interacts with the risk/reward analysis of the brain in a way which takes advantage of people’s irrationality.
“And it should be illegal. Allowing betting creates bad incentives.”
-Like what? Leading people to put their money where their mouth is?
The simplest objection is that anyone who has enough money on the line will be taking active measures to ensure that they win the bet. Need a large terrorist attack in the next year in order to win a $20M bet? You could finance one and still come out ahead.
By that logic the stock market should be illegal because it creates the incentive to send assassins and saboteurs against all the competitors to the companies one has invested in. Building railroads should be illegal because sending rustlers, cutthroats, murderers etc against landowners along the proposed path might be cheaper than buying their property at fair market value.
Yes, I am aware of the railroad wars. That was one of the reasons I chose that example.
Now, should building railroads be as illegal as gambling, and for the same stated reason of “incitement to violence”?
Surely there are quite a lot of people already who value a victory by their preferred candidate at more than twenty million–most obviously the people who spend money on that scale supporting candidates.
Your observation leads to a rather sinister followup question:
Twenty million dollars would buy a very good assassin. In fact, half that would buy a decent one.
Why does nobody attempt to save some money and achieve the same result more directly and with a better chance of success?
Twenty million dollars would buy a very good assassin. In fact, half that would buy a decent one
I am reasonably certain that the elite, highly-paid killer for hire is almost entirely a creation of Hollywood. To the extent that there are genuinely professional assassins in the world, they are recruited from within the ranks of large organized crime groups and/or government agencies, and are not hired out. The demand for independent assassins with eight-figure fees and the supply of patrons with eight-figure bankrolls is too limited for there to be any market in which they could reliably connect with one another without being hopelessly outnumbered by con artists and police informants.
Several steps down the food chain, there’s a large supply and demand for low-level, usually non-murderous thuggery, and if you know a guy who knows a guy he can probably get a thug to step up to murder for $20K or so. But you’re not going to want to bet on that thug getting past the Secret Service, or keeping quiet when he gets caught. And he is going to get caught long before he builds his skills to the million-dollar-a-hit level.
Some things money really can’t buy. This may be one of them.
I seem to recall that the circumstances surrounding the arrest of Silk Road founder Ross Ulbricht involved him hiring hitmen to kill an acquaintance of his. (His Wikipedia page says he was charged with procuring murder, a phrase which I have never heard but quite like the sound of.) Having poked around on some .onion sites myself, I can attest that there are people claiming to be hitmen, and they aren’t terribly difficult to find. So they certainly exist, but I don’t know the odds that a given self-proclaimed hitman is for real.*
Anyway, in a communication between Ulbricht and these killers, he mentioned that they were charging him $150,000, but he was unhappy with the price since he’d hired someone before for only $80,000. So the going rate for murder is much less than twenty million. I imagine most people refrain out of a sense of decency. That and if you’ve got $20M to give, the money is probably worth less to you than avoiding the risk of your friends finding out you had someone assassinated.
*(The odds are probably not good. I thought about setting up a scam myself for would-be murder-procurers, to make some money and do the intended victim a favor. I didn’t have the energy, the resources, or any idea of how legal it would be, but I’m sure the idea has occurred to plenty of others who are more motivated and less law-abiding.)
A few tens of thousands of dollars seems to be the going rate for a freelance “professional” assassination, yes. It doesn’t actually buy much in the way of professionalism or loyalty, as Ubrecht found out. In cases where the assassins are caught, they turn out to be criminals who mostly make their living from less murderous crimes but were willing to up their game for a one-time windfall. Even organized crime syndicates often make do by just telling a kneecapper that he should set out to deliberately kill someone this time; only the largest and most violent organizations have enough work for a class of specialist assassins. And organized crime syndicates have the advantage of nearly birth-to-death tracking of their members, which helps weed out the police informants.
I have seen no evidence, from the supply, demand, or outcome side of the equation, that there exists a significant number of professional freelance assassins who will work for those outside their particular organization, will reliably kill specific victims without getting caught, and who are not mostly police informants. And I would expect there to be such evidence even if the best professionals never get caught and always make it look like an accident, because they can’t all be the best professionals. Unless we are dealing with a very bimodal distribution (and how do the pros learn their trade in that version?), it should be possible to estimate the number of never-get-caught assassins from the relative numbers of clumsy usually-get-caught-if-they-aren’t-informants-from-the-start killers and of skilled but unlucky freelance assassins who almost get away with it in spite of a reasonably professional effort.
If a class of such people did exist, and a market where they could be reliably contracted, I would expect from the prices of their amateurish colleagues that the professional killers would charge well over $100K per hit, so I’m not sure what Ulbrecht was thinking. And since when did the Dread Pirate Roberts have to hire freelancers, anyway?
Even if we had an underground Assassin’s Guild, they’re not going to want jobs where they don’t have a reasonable chance of getting away with it. So sure, you could probably take out a hit on some guy living in the projects and make it look like a drug deal or robbery gone bad. But taking out the candidate for a governorship or senate seat is going to bring down a lot of heat, such that the assassin is unlikely to be able to enjoy whatever fortune they charged you. Same for a large terrorist attack. It is very, very unusual that we end up not knowing who carried out political assassinations or terrorist attacks, which is why they tend to be done only by those who are ready to die or are completely insane.
The sad and puzzling murder of Dan Markel is one of the few cases I’ve read about that lead me to think that maybe there is a shadowy underworld of professional assassin ordinary people can hire.
But that aside, my stronger conviction is that trying to hire an assassin is more likely to end up with the solicitor in prison and the target alive than anything else.
While I take the points raised (and they are good) one possible semi-analogous situation is bank robbery. The FBI is fond of pointing out that they catch something like 90% of bank robbers and that the average bank robber gets away with only something on the order of a thousand dollars.
I once heard an argument which I found convincing that most of the other 10% are committed by a small cadre of professional bank robbers who are, essentially, smarter than the FBI and rarely get caught.
Full-scale high-tech-thriller-movie assassinations are indeed very rare, which is pretty strong evidence that high-tech-thriller-movie style assassins are likewise very rare. No question. However, I still think that with a seven-figure budget, the potential is there. I suspect that the primary reasons it doesn’t happen are a) fear of discovery (including betrayal) and b) fear of escalation.
The sad and puzzling murder of Dan Markel is one of the few cases I’ve read about that lead me to think that maybe there is a shadowy underworld of professional assassin
And Olaf Palme on the political front, but the numbers are too small to actually support a market. More likely, these are just occasional amateurs getting lucky.
As for the comparison with professional bank robbers, they have a huge advantage in that their crime gains them essentially untraceable cash. A freelance assassin’s crime gains them a debt from a stranger who is by definition a murderous criminal but who is not part of an organization that can verify his professionalism and loyalty. The professional bank robbers just need to exist in isolation, there doesn’t need to be an actual community of them. The freelance assassins would need a community, organized into a market that can match vetted professional and trustworthy killers with vetted professional and trustworthy employers or there’s no way any of them can expect to profit from the whole thing.
If you want an assassin you can trust to get the job done and rat you out, you’ll probably need to recruit and train him from scratch. I can see how that could be done for tens of millions of dollars, but not how it could be done in less than years, which pretty much rules out tactical electioneering by gunshot.
“A freelance assassin’s crime gains them a debt from a stranger who is by definition a murderous criminal but who is not part of an organization that can verify his professionalism and loyalty. ”
Looking over my bills this month –
homeowner’s association dues – better pay that or they’ll sue me
cable bill – have to pay that or they’ll cut off HBO before I get to see the new Game of Thrones
professional assassin bill – ok, this one I can skip – what’s he going to do? Sue me for the money?
>Also, gambling is an attractive nuisance that interacts with the risk/reward analysis of the brain in a way which takes advantage of people’s irrationality.
That is the entire point. Those who are prone to make decisions based on irrationality get a smaller chunk of the total pie because making the pie requires rational decision makers. It is only logical that the rational ones should have more of a say in how the pie is made, and thus a higher percentage of the pie. This is far from “bad incentives” and is in fact much closer to “good incentives” in terms of total productivity and output.
But the fact that irrational people lose money, money which goes to others, creates financial incentives to encourage irrationality. I count those as bad incentives as well.
30 World Events is higher than I would have gone, I’m curious if you’ve got an elaboration.
If I did my back of the envelope math right, are you suggesting that Clinton would only have an approximate 36% chance of winning the general election against any Republican not named Trump? That seems extraordinary, given that you think Hillary has an 80% chance against the Donald. That suggest to me that (a) I did my math wrong, (b) you are overestimating Trump’s nomination chances, (c) you are overestimating Clinton’s chances against Trump, or (d) you are underestimating Clinton’s general election chances.
My guess is some combo of (a) and (c).
Looks like 40% for P(Rubio Nominee) and 30% for P(Clinton|Rubio Nominee) so it’s 70% for P(Rubio President | Rubio Nominee). Also 100% for P(Trump Nominee ∪ Rubio Nominee).
Scott, do you think about predictions differently when markets exist? Do you consult the markets?
For example, the markets give Hillary only 80-85% chance of nomination; and they aren’t so negative about Trump’s chances in the general election, I think 50/50.
And you could probably back out from options any particular prediction about DJIA.
I tried not to consult prediction markets because then I wouldn’t actually be doing any work, just listing what the markets said.
Upon checking the markets, I’m sure they’re better than I am.
The markets aren’t far off from those figures, but those mostly aren’t markets, those are bookmakers, where you’re betting against the house rather than against other people, and you can only bet, not lay. They offer as short odds as they can get away with and still succeed in attracting clients, which makes them very rough (as the variation shows, market competition is there but imperfect). It’s pretty important to distinguish between bookmakers and markets because you need to factor this in when reading from them- in particular, you wouldn’t want to make the mistake of averaging the odds or similar.
To get a reasonable reading of implied probability from a bookmaker’s figures you’d probably have to do something like take the naive implied probability as an upper bound, then sum the naive implied probabilities of all the other candidates with non-tiny probabilities and subtract that from 1 to get a lower bound, and it’d be somewhere between. Those are the bounds you could make money by correctly assigning outside of.
Markets, on the other hand, will settle at the place where people are willing to both lay and bet, and should all have the same value where liquidity isn’t an issue. You can read off implied probability by taking the odds in the middle of the unmatched bets/lays from them relatively easily, and the gap will normally be a lot smaller than when working out implied probabilities from the distribution of a bookmaker’s bets.
You can usually trust Betfair as reference; it’s the largest, and any bookie longer than it or decently liquid market which differs from it in either direction will get arbitraged to it in short order (if the gap is greater than the 5% commission), because people look for arb opportunities pretty aggressively.
Mostly are quite reasonable. Just highlighting the few questionable ones (IMO):
6. Assad will remain President of Syria: 60%
Why so low? His position has arguably never been better since the insurgency began.
12. Syria’s civil war will not end this year: 70%
Extremely optimistic. More like 95%, I think.
14. ISIS will not continue to exist as a state entity: 60%
20. Hillary Clinton will win the Democratic nomination: 95%
Currently at ~80% on the predictions markets, though. Will stick by my original 90% assessment.
21. Donald Trump will win the Republican nomination: 60%
Approaching 50% on the predictions markets. I predicted 40%, but that was at the beginning of the year, when the predictions markets were actually giving him just 25%.
So yes this sounds about right.
31. Oil will end the year lower than $40 a barrel: 60%
Put that at 30% at the beginning of the year, and still sticking by it. Latest news is that analysts expect oil prices to be back up to $50 by Q4.
41* European far right makes modest but not spectacular gains: 80%
Hardly possibly for them to do so. Although 2017 should be interesting – elections in France, Germany, and likely a UK referendum on Brexit.
50. Occupation of Oregon ranger station ends: 99%
0.1% chance it grows into the National Secessionist Forces.
I predicted 60% on January 1 for the Donald winning the Republican nomination. He’s by far the most credible Republican candidate on foreign policy, and Cruz is an intellectual mediocrity. Trump seems to have what it takes to win. I put the chance of Clinton winning the Democratic nomination at at least 90%, close to 100% if both candidates stay alive and healthy. In Democratic presidential nominations, always bet on Black.
And, yes, Assad’s position is good. But the Syrian Army is really weak and can only make gains with firm Russian support, which is typically, but not always, effective and forthcoming. The Talbiseh pocket has expanded to a dangerous size since the Russian intervention, despite the Russian airstrikes.
Why are the analysts so bullish on oil?
And, yes, Assad’s position is good. But the Syrian Army is really weak and can only make gains with firm Russian support, which is typically, but not always, effective and forthcoming.
Well sure, though one can also point out that SAA itself appears to be getting better – recent photos have increasing numbers of them in modern body armor, and generalship is also improving judging by the recent operations to retake Rabia.
But even if it gets bogged down in another stalemate, why would it necessarily impact a lot on Assad’s political survivability (which is what the prediction was about)? Even if that happens, I still don’t don’t see how the chances of him keeping the Presidency this year could go down to 60%.
The Talbiseh pocket has expanded to a dangerous size since the Russian intervention, despite the Russian airstrikes.
I don’t recall any big recent rebel expansions there. I just heard of some RuAF air strikes on it and that’s all. Can you provide a link to a before/after map? In any case the pocket there is far smaller than it was in 2013 or even 2014. I’m not sure how it can pose any sort of threat now.
Why are the analysts so bullish on oil?
My guesses would be:
(1) Rise of US interest rates have already been priced in;
(2) Cost of marginal global oil production is substantially above current low prices, so high cost producers are reining back in. The sharp rise in US oil production, enabled by new oil technologies, that was observed over the past few years has stalled and has in fact been declining again since mid-2015; a decline that they predict will now accelerate. So one can expect the current glut to pass and oil prices to go back up to equilibrium.
(3) The Chinese continue to buy 20 million+ cars per year.
“Even if that happens, I still don’t don’t see how the chances of him keeping the Presidency this year could go down to 60%.”
The more energy-intensive industries in China aren’t looking good in the short run. In the long run, yes, Chinese consumers will have a huge effect on oil demand.
And the Talbiseh pocket has expanded northward and eastward, almost reaching Hama city. That was (as far as I can remember) never the case before late 2015. Mid-2014:
After mid-2014, the Template largely ceased to be modified in favor of the Module, a less visual representation:
I can see everyone getting sick enough of all this that the Russians broker some agreement where Assad steps down in favor of someone else exactly like him and everyone declares that Justice Is Served. I’m not even sure Assad would be too opposed to it – he never really seemed to want the job that much.
Regarding far-right gains, I’m willing to count non-electoral gains (eg up in polls, more positive coverage) if no elections.
Okay, fair enough.
There have been reports that a general went to Assad with an offer from Putin to step down but it was refused (however Russia itself denied that happened).
In his various interviews with the foreign media Assad has to my knowledge been adamant he will not leave while Syria remains in civil war, except via elections, arguing that it would be irresponsible to do so. So, prima facie, that’s a no.
Up in polls sure, more positive coverage – not sure at all. Many European countries seem to be clamping down on nationalists and the far right, e.g. AfD has been blacklisted from TV debates by the German state news broadcaster.
I suspect Assad wants very much to not be lynched, assassinated, executed, or imprisoned for the rest of his life – and in the age of the ICC and universal jurisdiction, where other than the Presidential Palace in Damascus can he reliably avoid at least the latter of those fates?
Russia certainly could protect him, but how reliable would such a promise be once Assad is no longer in a position to do any favors for Putin? A little polonium in the tea, to make sure there’s no interference with the new puppet, and a Russian hospital to certify death by natural causes…
Scott, have you ever looked at Assad? Clearly, he is the most West-palatable non-Sunni person in the whole Syrian government. I don’t think either the West or Russia is that stupid. And see Schilling’s point.
“I’m not even sure Assad would be too opposed to it – he never really seemed to want the job that much.”
-From where did you get that impression? Assad claimed the opposite in his 2011 interview with Barbara Walters.
The ‘expert take’ on Assad stepping down seems to be that he’s bitterly opposed not out of desire for power, but fear of revenge. Syria has a fairly ugly history of genocide even before the current radical Islamists are factored in, and the Alawites as a whole would be at real risk if he stepped down in a Libya-style non-solution.
I think Assad would probably be willing to hand over the reins at this point, but only with some guarantee of a strongman government capable of preventing the backlash he’s invited.
You list both “18. No country currently in Euro or EU announces plan to leave: 90%” and “43. Nobody leaves the Euro or EU this year: 90%”. Unless you believe that p(a country leaving the Euro or EU|a country has announced that they’re leaving the Euro or EU)=100%, I’d call shenanigans. If you do so believe, I disagree with your assessment but withdraw my objection. Also, it might be a good idea to explicitly indicate that asterisks mean that they’re subjective for the benefit of those who’re new to this blog.
21. Donald Trump will win the Republican nomination: 60%
25. Marco Rubio will not win the Republican nomination: 60%
== Marco Rubio will win the Republican nomination 40%
Can I bet my $0 against your $1 that someone other than Trump or Rubio will win the nomination? I’m fairly confident it’s > 0% chance 😛
There are similar inconsistencies in several other probabilities, though more complex (e.g. trump * trump lose = 48% dem win (* 95% hillary). Then, for the 40% not trump nomination, hillary clinton most likely loses? (60 hillary win – ~45% from trump scenario = ~15% of ~40%, so she loses 5/8 times to rubio?)
“18. No country currently in Euro or EU announces plan to leave: 90%”
Given that the UK is due a referendum on EU membership and Remain/Leave are basically deadlocked in the polls, this looks heroically optimistic.
Agreed. I placed probability of Brexit at 50% in my predictions (linked to at the bottom of Scott’s post here). But I placed the probability of Eurozonexit for any country at 0%.
Yeah, this is the one I came to the comments for.
There’s still a possibility the referendum will be sprung in the UK this summer, and if anyone can call the referendum result with 90% confidence I can only bow in wonder at their prediction skills.
(Then again, I hold Scott in such high regard that his “bitcoin goes way up” prediction at 80% had me seriously considering making some unwise investments ;p)
Eh. Betting markets give ~67% chance of staying in a referendum conditional on one being held (decimal odds convert to implied probability as 1/[odds], I believe, although on politics, away from major election frontrunners, liquidity is weak and you could plausibly even manipulate it), so 1 – (0.33 times probability that the referendum is this year)… *after* having looked at the markets I’d call 0.9 for no announced departure this year optimistic, but not too vastly so.
I think it’s good for not having cheated ahead of time.
Any comment on the two order of magnitude discrepancy in the earthquake measures? I’d consider an earthquake with 1000 deaths to be pretty major (in line with this result after searching duckduckgo “earthquake casualties”.
Depending on the location, the death count can vary a lot, e.g.
2010 Chile Earthquake: Magnitude 8.8, 550 deaths
2010 Haiti Earthquake: Magnitude 7.7, >100’000 deaths
the difference being due to Chile being prepared for large-magnitude earthquakes.
That said, the numbers Scott gave seem pretty arbitrary. The prediction is a weird compound between the location of the earthquake, how usual earthquakes are in that location, and how rich people living there are. If the focus is meant to be on the human cost, why make predictions about earthquakes specifically instead of natural disasters in general?
With regard to 1, does the U.S. putting substantial ground troops into fighting ISIS count as a “new major war”? supposing there are substantial U.S. casualties?
With regard to 30 …
Bitcoin is currently a little below $400. If you buy a dollar’s worth today, 80% chance of $500 gives you your dollar back. In addition there is the expected value of the 20% chance of values under $500 plus the expected value of the price over $500. Hard to believe that doesn’t come to well over the current interest rate.
So curious minds want to know—are you betting on your prediction?
Exactly what I wanted to ask (the bitcoin part)
I took over a blog last year, and I’ve made some 2016 predictions modeled after your prediction posts. It’s a pretty small blog and there’s only about 20 predictions, but if self-promotion is allowed, I figured this would be a good place to post them: http://postlibertarian.com/2015/12/31/2016-predictions/ My name should also link to the post.
I have made predictions on a few of the same events Scott has, and we differ in certainty levels mostly. The only big difference in direction is that I had Donald Trump not getting the Republican nomination at 80%. I did write these at the end of December, and if I redid them today I’d probably give him a better chance to win, but not as high as Scott.
Several of these predictions are quite uninformed. For example #33, the Shanghai index, is already down 18.6% since it opened this year. That is well past 10%. It’s a bit bold to think it will rebound > 8% before the year end, considering the weak state of the Chinese economy.
On #6, you give Assad a only 60% chance of remaining in power? Assad is more secure this year than most. America will be unable to sway the battle in the region much due to the pressures of an election year, and Russia is openly backing Assad. I’d give Assad a 95% chance of remaining in power.
Again on #9, the chances of progress with Gaza or peace negotiations approaches 0% in 2016. Not only is Israel less interested than usual (thanks to all the Palestinian knife attacks), America is in an election year. Obama has grown more distant from Israel, decreasing the chances of any peace overtures, and even early peace talks would prove to be easy fodder for Republicans given the current anti-islamic mood of the nation. Obama won’t bother, it could only harm the Democratic nominee, so there is nothing new coming down the pike from America. Europe is too embroiled in their own migrant troubles to notice Israel. Iran is busy re-entering the world stage. There are no interested parties for peace in 2016.
Regarding #47, the California drought, it would take at least 5 more years of this current El Nino for California to return to “normal” conditions. Considering the technical challenges in compressing 5 years into 2016, and barring any advances in time dilation or time travel, there is a 0.01% chance of the drought ending this year. Not 50%. Reservoirs remain extremely low, and only have barely begun trending upwards. http://cdec.water.ca.gov/cdecapp/resapp/resDetailOrig.action?resid=NML
Hey, we don’t know Scott’s financial situation. For all we know, Scott blinded the car dealership with science and is paying less than market rate in interest, and has invested the difference.
I don’t think this is likely (…10%?), but I also don’t think that we have enough information to finger-wag at this point. A short-term lease which moves directly into buying a car or walking away trades a little extra payment for a lot of extra information. Since we don’t know the terms of the lease or the properties of the car, we can’t judge the effectiveness of the use of money.
Clearly the answer here is for Scott to post, as a supplement to his yearly predictions, his yearly income and spending, so that the Internet can helpfully optimize it for him.
@ The Voracious Observer:
Your link shows a particular reservoir that is low. For figures on a bunch of reservoirs, see:
If I read it correctly, that shows total storage at 56.55% of its average level, I presume for this date. I wouldn’t count on the drought officially ending this year, but your five years looks like a considerable exaggeration.
I’m not an expert, but the figure I heard was five years of normal rainfall, or two or three years of a strong El Nino. This is shaping up to be a strong one, so another year or two of this would probably put us at parity.
What do you base your judgement of this rainy season on? From what I’ve seen, we’ve had plenty of cold but scarce rain. I’m hoping the mountains have gotten more but I don’t know.
Snowpack was at about 130% of average last I checked, which was a couple weeks ago, and predictions for the season were that it’d get wetter later.
If the rainfall pattern in California continues to look like a typical El Niño year, the drought will be over. The water shortage may not be, though.
“The drought is over” has, I think, two different meanings.
1. We are not getting an unusually low amount of rain and snow this year.
In that sense it looks as though the drought is over.
2. The reservoirs are now back up to their normal level.
In that sense the drought is not over, and probably won’t be by the end of the rainy season this year.
I think 1 is what the phrase sounds as though it means, 2 the way in which it is mostly being used.
The end of the water shortage is going to require a change in human institutions, laws mostly, not just more rain. Given California’s dysfunction I wouldn’t hold my breathe.
And worse, as was pointed out in Scott’s post about this, it will require changes to state *and* Federal laws, in ways which will be challenged in courts for decades.
I’m not familiar with the operations of that system. Do they use a rule curve like we do in the PNW?
Reservoirs are normally drawn down in the winter and filled with the spring freshet here.
Well, this post inspired me to subscribe to the Unsong subreddit, so there’s that.
I really admire the whole exercise of creating prediction lists like this. Definitely a good way to improve your thinking.
Mostly pretty reasonable. The biggest disagreement is I think you’re drastically overestimating Hillary’s chances of winning the Democratic nomination. I’d say we’re moving into a situation where Sanders is the odds on favorite, but even if you’re significantly more pessimistic about his chances than I am 5% seems absurdly low.
I’m surprised you think there is a 40% chance of more than 100 deaths in a US war this year. That’s pretty pessimistic.
> US will not get involved in any new major war with death toll of > 100 US soldiers: 60%
I say 92% chance.
Regarding bitcoin, are you familiar with this: https://medium.com/@octskyward/the-resolution-of-the-bitcoin-experiment-dabb30201f7
I’ve been considering doing these. I was thinking it might be interesting to do predictions for someone else’s personal life, and see how accurate those prove. I figure that it being one’s own life might make one biased in certain ways that wouldn’t happen when predicting someone else’s. Of course, I don’t personally know Scott, and have a very fragmentary picture of his life, so it’s hard to say which would dominate. If I do mine, I might also put down predictions for Scott’s life events.
As usual, most of these are pretty good, but some significant disagreements and why:
#3: Last year you had 60% for Greece staying in the Eurozone, now 95%? None of last year’s factors have been resolved that I can see, just kicked down the road (and out of the headlines) a bit. Probably by a few years, but I’d put more than a 20% chance of their reemergence this year, maybe due to the added financial strain of the refugee crisis, and that gets us to at least a 10% chance of an announced Eurozone departure.
#5: 80% is a nearly perfect status quo prediction, with 100-casualty first-world terrorist attacks happening once per five years in this century (up from once per decade towards the end of the last). I think the fallout from ISIS and the Syrian civil war is likely to double that risk, so 60%.
#14: 60% for the effective destruction of ISIS in one year? None of the people actually fighting ISIS have committed nearly enough effective ground troops to take/liberate all of ISIS’s cities, that quickly, and I don’t see anyone lining up to send more. 60% for ISIS being clearly on the ropes and as doomed as, e.g. Nazi Germany in 1944, sure, but very little chance of their not holding some territory. Call that 20%.
#18: 80% for nobody announcing a plan to leave the EU is optimistic for about the same reason as #3 w/re Greece and the Eurozone. The divided British electorate will probably grudgingly accept staying in the EU as they get closer to the referendum and start looking at the real costs, but I’d put the probability of an announced British exit at 20% alone, so make it 70% for the announced stability of the entire EU.
Which is not to say 30% chance of EU collapse, as an announced departure is as likely to be a negotiating position as a firm commitment. And the Union could well survive a select departure or two, for that matter.
#20: The Democratic nomination is still Hillary’s to lose as it always has been, but six months ago it was “unless she has a stroke or is convicted of a felony”. All the usual rules are suspended this year, the perception that Trump is now a viable candidate has spilled over to Sanders, and there are a fair number of Democrats who don’t really like Clinton. The level of scandal or misstep that it would take for Hillary to give it up to Sanders is now I think well above the 5% longshot you are still calling it. I’ll give her 85%.
#21: And I’ll give Trump 25%. No matter how much the media wants to sell the story that Republicans are crazy enough to nominate The Donald, he has only 35% of the vote – and maybe 5% of the party elite. The entire protest block is pretty stable at 45%, and some of those are evangelicals who will support Carson but not Trump. Arrow’s Theorem and a multiparty nomination notwithstanding, you do not win an election against competent opposition when most of the electorate wants Anybody But You. And weak plurality wins in the early primaries are not the bandwagon that will carry anti-Trump voters into his camp.
#25 was dumb but I almost agree with it given my take on #21. Call it 25% Trump, 40% Cruz, 35% Rubio. Though the longer Trump stays viable, the worse it will look for Rubio (I think).
#32 and #33, unreasonably optimistic given events to date. I’d put it at 50% for the Dow, 30% for Shanghai.
#35: 95%, not 99%. The North Korean leadership is not nearly as crazy or stupid as they are often made out to be, but they are never 99% certain to steer clear of a major crisis and most of their major crises involve coastal waters filled with tiny stupid islands.
#36: 80%, not 95%, same reason.
#45: 10% if you mean orbital launches; there’s no room for that in SpaceX’s 2016 manifest and it would be an expensive stunt to try and slip one in without a paying customer. Elon’s been pretty consistent about waiting for customers to share the cost before he does stuff like that. Suborbital tests, yes, 50% would be about right.
#46: 95%, though it depends on what you count by “nobody important”. At this point, anything that would persuade an EMdrive skeptic would constitute experimental demonstration of a perpetual motion machine. I am 99% certain that this won’t actually happen and 95% certain that nobody important will think it has happened. Also 99% certain that no important EMdrive believer will change their mind over a year of null results.
Scott’s personal life:
No opinion, but I am glad that you are so confident in the future of this blog. You’ve created something wonderful here, and I’d like to see it endure with a ~20 year projected life expectancy (yes, I know that’s an oversimplification).
And I do wish you could be more confident of your housing, romantic, and professional success or at least stability 🙂
“but six months ago it was “unless she has a stroke or is convicted of a felony”.”
-And it’s exactly the same today. Nothing’s changed.
“most of the electorate wants Anybody But You.”
-Not most of the Republican electorate. Most of the general electorate, yes, but that’ll be fixed after the nomination.
“60% for ISIS being clearly on the ropes and as doomed as, e.g. Nazi Germany in 1944, sure”
-Doubt it. As you know, I think Obama created IS (and killed Bin Laden to take control of the global jihadist movement), so it’s strongly doubtful he’ll be giving up that asset. That would require letting the Axis of Resistance win. Unlikely the U.S. would ever allow it to do that.
I agree. Even if a miracle happens in Iraq and Syria, ISIS still has big chunks of territory in Libya and Yemen.
Libya and Afghanistan. It has as much territory in Yemen as it has in Lebanon.
Right. I remembered seeing a map of the Yemen Civil War showing a big chunk of ISIS territory in white, but then I checked again and it turns out that that was Al-Qaeda’s Yemen branch, not ISIS. My mistake.
The franchises aren’t ISIS is the relevant sense. If ISIS proper gets kicked out of Syria the Powers That Be aren’t going to be able to waltz in and take over in Libya.
Just wondering if you happened to see my post about the process of scoring the results of these predictions?
What does Cast Lead mean? (points 8 and 10 in world events)
Scott, may I ask why you plan to vote in a clearly non-swing state?
I’m naturally curious about this since a big part of what’s written here is inspired by SSC and LW.
Yeah, I read that post of yours and strongly disagreed with it.
You could say the same thing about any coordination problem. Suppose you’re in a game where 10000 people have to put one penny into a pot; if greater than 5000 people do so, everyone wins $1000; if fewer than 5000 people do so, nobody wins anything. Your penny has almost no chance of changing the eventual outcome, so you recommend not giving the penny? I would say the opposite – giving the penny is morally good, and you should use Kantian “I will pretend my action determines everyone else’s action, even though it doesn’t” type reasoning.
This makes even more sense when you think of it in terms of demographic groups. Say that you successfully create the norm among rational econoliterate people that voting is stupid and you shouldn’t do it, and that (really REALLY stretching it) these people are half of the population. On the other hand, some televangelist or whatever tells the irrational illiterate half of people that God wants them to vote, and so they do. Well, now all political power belongs to irrational illiterate people, and further, they deserve it, because we spread stupid counterproductive ideas and they spread good ones. This seems like a clear case of “rationalists should win”.
Even as I was reading your first paragraph, I was thinking “Maybe the propagandizing is more important than the penny”, and your second paragraph may point in the same direction.
A penny is cheap enough. Voting has a somewhat higher time cost, unless the UI for contributing pennies is remarkably awful.
I’m not sure how important it is to actually vote vs. convincing other people to vote in your preferred direction.
While we’re at it, what do folks here think of “I don’t care how you vote, just vote”?
I endorse it because it minimizes the chances that any particular faction will actually accomplish anything. People accomplishing things rarely goes the way I want it to, and when it does, it usually turns out I was wrong to want them to go that way.
As far as its moral assumption, I find it ludicrous.
I think there are three different arguments here, one of which is wrong:
1. The total benefit of getting the right person elected is so large that, including benefits to others as well as to me, it’s worth voting for even the tiny effect my vote has on the probability that the right person will win.
That’s a legitimate utilitarian argument. If we guesstimate one chance in ten million of a one vote victory (ignoring lots of complications due to the electoral college system), and believe that your candidate winning is worth $100 each to all Americans, that’s a $3000 benefit, so worth the trouble of voting.
Of course, you ought to discount it for the probability that you are wrong about which candidate is better. After all, the only reason the outcome is in doubt is that about half the voters disagree with you.
2. Arguing against voting may convince enough people with views similar to yours not to vote to have a significant negative effect on the chance that your preferred candidate wins.
That’s a legitimate argument, but it’s an argument for keeping quiet, not for voting.
3. The Kantian argument–that you should act in the way you would like everyone to act. That makes no sense to me. My voting has no effect on how other people vote, so I should decide how to vote in terms of direct costs and benefits to me plus costs and benefits to me and others whose welfare I value of the effect of my vote on the outcome of the election.
Perhaps Scott would like to offer a defense of the position? I can’t see one. Presumably, if he’s at sea, his principle forbids him from moving to either side of the boat.
On that last one: The trick with the Kantian approach is always, always, always to figure out exactly what it is you are trying to model. (For instance, I once read a very well-argued essay using the Categorical Imperative to make a moral case for chattel slavery.)
If everybody voted based on their “true beliefs,” or whatever it is that should motivate the vote of a thoughtful person, then democracy would work at an optimum as an expression of the will of the populace. So if you want democracy to work as an expression of the will of the populace, you should vote accordingly.
On the other hand, if you think that everybody should vote according to a cynical calculus incorporating how close you can get to what you want while still supporting a candidate who has a chance to win, then you should do that.
And if you think that everybody should refrain from voting because its costs outweigh its benefits, then you should do that.
That’s the wonderful thing about Kant: he can lead you down any path you already wanted to follow, profoundly and with great confidence.
“If everybody voted based on their “true beliefs,” or whatever it is that should motivate the vote of a thoughtful person, then democracy would work at an optimum as an expression of the will of the populace. So if you want democracy to work as an expression of the will of the populace, you should vote accordingly.”
Do you mean “you should if you accept the categorical imperative”? The question I was raising was why one should do so.
Absent the categorical imperative, your conclusion only follows if you believe that your acting that way will make everyone else act that way, which it obviously won’t.
Yep, that’s what I meant. And you’re quite right: if you don’t accept the CI, then the whole thing is just ridiculous.
(more specifically, in the boat example, universalize the principle you’re actually using – “I go to whichever side of the boat I want”. If everyone wants to be on the right side for some reason, then you actually do have a problem)
The Kantian argument can be gotten around by just voting with some trivially small probability. In fact, only some random subset of the votes get counted, so essentially that is already the system we use.
So your objection to non-voting rationalists is I can imagine a world in which voting mattered?
I mean, sure, if rational econoliterate people were half of the population AND they were all completely agreed who the better candidate was AND the populace-at-large hadn’t reached the same conclusion then it might make sense for rational econoliterate people to vote. But in the world we actually live in:
– candidates say whatever they think will appeal to their voters or swing voters
– candidates lie about their own views
– candidates change their mind after being elected
– elected officials discover their actions are constrained (by politics, by secret information, by changing circumstance) such they CAN’T do what they said they would do, or can’t do it well enough that it would remain a good idea.
Which all means the fact that a candidate says they would do X is just about the weakest possible evidence that the candidate would do X.
So, on what basis are rationalists supposed to be deciding who the best candidate is? What makes you think our pick would be better than that televangelist’s pick?
If I vote, I can be ABSOLUTELY SURE it will make me dumber about politics, I can be ALMOST absolutely sure my vote will have no effect, and can be pretty sure that even if my vote DID have an effect I don’t know whether the effect would be positive. So the math says I shouldn’t vote. Doesn’t it?
“If I vote, I can be ABSOLUTELY SURE it will make me dumber about politics, I can be ALMOST absolutely sure my vote will have no effect, and can be pretty sure that even if my vote DID have an effect I don’t know whether the effect would be positive. So the math says I shouldn’t vote.”
I vote in some years, even though I agree with all of this (subject to the caveat of replacing “absolutely sure” with 99+ percent sure). Offsetting the effect of becoming dumber about policy are a) the pleasure of actively rooting for my team (the same reason I follow sports), and more importantly, b) an increased sense of civic connectedness, making myself less misanthropic towards my fellow citizens. These personal effects vastly outweigh any effect I can have on outcomes. Of course, the risk I run is that the years when I’m the most stupid about politics are likely to be the years when I’m least likely to recognize that fact, and thus get the calculation wrong about whether I should vote this year.
When I do vote, I sometimes write-in non-candidates for the top offices (President, governor, senator). I don’t mind throwing away my vote, because I assign negligible probability to affect outcomes, and am only voting for the effect on myself. I should probably write-in names more often.
Ah. I’ve never understood that either. I think if I grokked why people care whether “their” sports team wins to the point of “rooting” for a team, political involvement might make more sense than it does.
In the past when I did vote it was for libertarian candidates who never won and rarely got even a double-digit vote percentage. Thus, my experience has always been that voting gives me a decreased sense of civic connectedness – an increased impression that the voters at large disagree with my views. (This is somewhat reinforced by the fact that every time I see an ad AGAINST a candidate I find it somewhat persuasive in the opposite direction as intended – it makes me like the demonized candidate more than I thought I did.)
So are you saying you experience an increased sense of connectedness when you vote for the winning candidate? (I suppose that might be true – I’ve never actually tried voting for the winning candidate.) Or was it the voting itself that did that for you?
As it is, I have come to regard national elections as primarily a form of entertainment. Given that premise, Trump versus Sanders seems like the best primary result…
Why is it that sports teams, unlike most other firms, have a clear identification with either a city or a university?
I think the answer is that part of what they are selling is the pleasure of partisanship, of rooting for your side. That link gives them a pre-existing body of partisans.
Every four years a game is played out across the nation with the fate of the world at stake. You can not only cheer for your side, you can play for your side, admittedly in a very minor role, for the cost of half an hour of your time. Who could resist?
My old solution to the puzzle of why, given the argument for rationally ignorant voters, anyone bothers to vote.
“So are you saying you experience an increased sense of connectedness when you vote for the winning candidate? … Or was it the voting itself that did that for you?”
The effect is larger when I vote for the winning candidate, but is still present when I vote for a loser.
I think that David Friedman is correct – rooting for sports is essentially a form of voluntary tribalism, with all the attendant benefits. You get the feeling of belonging, an adrenaline surge when your guys win, and a support system for when your guys lose.
The geographic base is another good point – I was a very lukewarm Royals/Chiefs fan up until a few years ago, when I moved away from Kansas City. Now I find my affection for both teams has massively increased, because by rooting for them I proclaim my affiliation to my home (and my extension my kin there).
I’m gonna gloss over the “rationalists” bit because this applies just as well to non-rationalists, but given your premises there are still several reasons to choose one candidate over another:
– You could vote for the guy whose ethics seem closest to yours; it’s only imperfectly possible to conceal them under the kind of spotlight you get as a candidate. (This is what most independents do. It’s what most partisans do, too, but their ethics are entangled with partisanship in a way that makes the choice trivial.)
– You could vote for the smartest guy, or the best negotiator, or the most likeable personality. This is noise if you can’t accurately judge domestic policy, but it is important on the international stage.
– You could vote for the guy you see as the least, or the most, likely to go off the reservation, depending on what you think of the status quo. This has to be considered in light of secret information and the constraints of the office, but I think your relative perceptions are likely to be accurate. (This is going to be more important in this election than in the last few, and I suspect it’ll get moreso in the next.)
tl;dr don’t vote the issues, vote the man. Or woman in Hillary’s case.
That’s fair where outcomes are in doubt, but for non swing state voting the better analogy would be that every single one of the 10,000 people has to put the penny in, at which point you probably don’t get the payoff anyway. (though tossing it in with the local elections doesn’t cost a penny either).
Oh, don’t worry. Each year I generate a random number between 1 and 1000 and vote only if the number is 1. Following Kantian reasoning, if a random 0.1% of people voted, everything would be fine, with only a trivial increase in noise.
Further, if I don’t vote, I donate a small sum to my favored candidates. Presumably, they use this small sum to pay for TV commercials that would make me gag, but influence a large number of mouth breathing voters — a vote only costs about $30. I think this addresses your concern about demographic groups.
There are non-presidential elections that have much smaller margins. You’re also assuming that 50%+1 is equal to 51% or 55% or 65% and that bigger victories don’t generate more political capital.
I’m just here because you forgot the tl; dr version of your own link:
The marginal communicative value of a vote (vote total++, your candidate++) is larger than probability adjusted value of swinging the election. Still small, but larger, and so your analysis should include or perhaps even focus on it.
Frankly I do it as self-signaling; I do a lot of things for that reason.
My conscious consideration of decisions is imperfect – I find I make a lot of them using the heuristic of “what would a person like me do?” But my introspection is also imperfect, so I rely at least partly on my own past actions as a guide to what kind of person I am.
Thus it becomes important, when possible, to act like the kind of person I want to be.
In the case of voting, I vote because it’s a community expectation and I am intentionally self-signaling that I am a responsible citizen of my community. Hopefully my future heuristic autopilot decisions will be more responsible and community-minded in consequence.
In this age of proxy warfare, aerial bombing, and drone strikes, does it still make sense to classify a “major war” as one with a certain number of own military casualties?
31. I will not be Chief Resident next year: 60%
Growth mindset, Scott! 🙂
47. I don’t manage to make it to my friend’s wedding in Ireland: 60%
Depending where your friend is getting married, what time of the year, and how badly you need cheering up, that’s a pity. Then again, you may not have the spare time available to lose a week of your life at a drunken hooley. (Is it a country wedding?)
I’m not sure I want to be. It’s a prestigious position, but the prestige can’t actually be translated into anything useful (I think there’s an undersupply of psychiatrists, that it shouldn’t be too hard to get a good job, and that the jobs that are selective don’t select on whether you were chief resident or not) and it’s very unpleasant and a lot of exactly the type of work I’m worst at (taking lots of phone calls and organizing things).
Probably what will happen is that our boss will ask who wants to be chief resident, me and all my co-workers will stare off nervously into space, and finally one of us will have their urge for prestige overcome their common sense, break, and say “I guess I could do it if nobody else wants to”. 60% chance I’m not the one who breaks.
I can sympathise on the not wanting to deal with people thing. Chief Resident sounds like the kind of thing you get advised (and pressured about) when you’re getting careers counselling: it will look so good on your CV! You need to show that you can achieve leadership goals! etc.
I should have read the prediction about “I will not get drunk this year” before the bit about likelihood of attending the Irish wedding 🙂
It’ll be good craic!
Also there are definitely readers of yours attending (not just me) so you can write it off as an SSC social gathering on your socialising returns.
I really want to, but the realist in me points out that it’s two twelve-hour flights, plus another day or two of landbound travel time each way, to go to a place where practically the only person I know is you, and you’re going to be so busy getting married that this is probably literally the worst time all decade to visit Ireland in terms of opportunities-to-spend-time-with-Eoin.
I’m an Army officer who came out of company command last year, where I was in charge of 150 people and in charge of solving their problems. Now, I’m working as a structural engineer for the US Army Corps of Engineers, where my workday is to take a pad of paper and a calculator and solve engineering problems.
I love my current job so much more than my previous one. Being in charge of people is a separate skill set, and I find it to be a massive pain in the ass that’s not worth the psychological benefit.
To be fair, physical problems are more uniformly responsive to the same set of input than are human problems.
Solving human problems for a group of n people is like surviving falling off a mountain n times, adjusting for path, time of day, and velocity as you go.
The people who are good at it have my respect – except when they come onto my turf and expect equal flex in solving the problems I fix.
Wouldn’t it be trivial to get 100% accuracy on these predictions by making a lot of very certain ones? e.g.
The sun will not explode: 99.9999%
Mercury will not explode: 99.9999%
This works for any percentage: E.g. if we’re making 80% predictions, we can just flip the truth value of every fifth statement:
The sun will not explode: 80%
Mercury will not explode: 80%
Venus will not explode: 80%
Earth will not explode: 80%
Mars will explode: 80%
It seems to me that the number of predictions you get right are only meaningful when you compare yourself to other people assigning percentages to the same set of predictions. But then it just becomes a test of knowledge, not calibration.
Yes, but then it would be stupid. It would be like a weatherman saying “Either it rains today or it doesn’t.” You’d always be right, but what’s the point?
Assume that Victor has an intrinsic motivation to be considered “right” and uses a more sophisticated version of Vitor’s scheme to indeed _be_ “right”, how would you differenciate Victor from a person that is always right?
Now if Victor converted to the rationalist camp and wanted to test his calibration, taking his good intentions for granted, how could Victor himself make sure that he is indeed testing calibration and not “being right”?
And now, by analogy, assuming that in reality, we all, including you, are somewhere between the original Victor and the perfect rationalist, what steps did you personally take to not be prone to this problem with your post?
Exactly this! I couldn’t have said it better. The whole rationalist thing is about mistrusting people’s judgements, especially when they judge themselves, and designing mechanisms that work even when stupid things like this are done.
My example is the purest, most extreme version of what can go wrong. Whether it’s by oversight or dishonesty doesn’t matter; I trust you, Scott, to do this honestly (but still not perfectly without bias), but will the people who are inspired by your yearly posts do likewise? If a meme spreads that this is “the rational way of having opinions”, isn’t it wrong to implicitly give a method which has a proven vulnerability the rationalist seal of approval?
There probably is a mathematically sound way of doing calibrated predictions, and we should do our best to find it (whether it is in the literature or as of yet undiscovered) and adhere to it.
This form of gaming seems easy to spot – look at the actual predictions they made and see if they look fishy (e.g. lots of 99% or 50% predictions). Don’t just stop at the calibration score.
I’m not sure there’s a mathematical way to prevent this, because you’re not manipulating the statistics, you’re manipulating the data you feed into them. Every statistical method in the world is vulnerable to cherry-picked data.
However, what you could do is test your predictions against someone else, and see who is better-calibrated. That way, any gaming of the questions should benefit both of you equally.
Scott’s method of scoring his predictions does suffer from these problems that you mention, yes.
Many other methods of scoring predictions exist that avoid them, though. Basically, one picks a scoring rule which calculates (given the outcome and the probability assigned) a score for each prediction.
You almost certainly want your scoring rule to have the property that your optimal choice is to choose the actual probability that you believe. Even demanding this property leaves a few good choices. A common favorite is to score the logarithm of the probability assigned to the actual outcome (these are negative, so you might want to normalize by giving log (2) points per prediction so that you can get 0 points by picking 50%. Obviously the base of the logarithm doesn’t matter; 2 or e are good choices I guess?
How useful is this kind of predicting? Certainly it’s honest to give a confidence when doing predictions, but I wonder how meaningful it is when doing so many predictions.
It seems to me you’re basically doing something related to a Fermi approximation here. Some predictions you’ll be too confident, others not confident enough, but over so many predictions that’ll probably cancel out and you’ll always end up with something like “I did 100 predictions with average 70% confidence, and got 75 correct. That’s pretty good”
Your bias would have to be HUGE to be significantly distinguishable from random noise. With so many predictions weak predictions over so many different subjects, it seems exceedingly unlikely that anyone would have such a huge bias.
I think to be meaningful you have to make much stronger predictions. In the 99% or 99.9% confidence ranges. Make a 100 predictions with 99% confidence and at the end of the year we can see if you got more than 3 wrong and say something about your predictive powers.
Right now you have 100 predictions with 73% confidence on average. Your 2 sigma interval would be somewhere from 64 to 82 correct predictions (I think, didn’t do the exact math). That’s a pretty big interval.
Put simply: It seems to me that this lacks statistical power.
Yeah but if he gets half his 99% predictions wrong that tells you something different than if he gets half his 50% predictions wrong.
If you want a larger sample size, well, this isn’t the first year Scott has done this.
I don’t want a larger sample size! A larger sample size just makes the results even more meaningless! Successfully predicting that a coin will land heads roughly 500 out of 1000 times is not more impressive than predicting it will land heads roughly 50 out of 100 times.
The only way to make predictions like these meaningful is to make them bolder. But boldness can only be established by comparing it to other people making sufficiently similar predictions. If all experts predict the stock market will go up, and it does go up, then that’s not very interesting. But if instead one expert says it will go down, and it does go down, then we’re on to something.
But most of the predictions on this list are fairly mundane, things most people will agree on. And they are also very diverse, covering many different topics, with confidences generally in the middle of the range. If I let a 100 people look at these predictions, I think their confidences will differ for many individual points, but their aggregate confidences will probably be pretty similar. Which makes the list kind of useless.
You should try doing Black Swan predictions. Rank events by expected (qualitative) impact, along with a guess of the attached probability. For example, what is your assessment of the probability of a nuclear terror attack on American soil? If the probability is > 1e-3, that event probably has more expected impact than a lot of other things (eg who wins the presidency). What is the probability of another financial crisis in which more than one large financial company goes bankrupt? What is the probability of an epidemic disease outbreak with > 1e5 victims?
Why should Scott work on Black Swan events? They’re really hard.
The interesting thing might be events (like the financial crash) which have strong precipitating factors that most people are ignoring.
I find it interesting that you make general political predictions, and some personal/social ones, but you haven’t posted any here in your field of expertise. (I mean, the janitor could have made the same sort of predictions wrt the work place, calibrated to his/her work team.)
What of the availability of drugs for treating your patients? What about an increase or decrease in the US suicide rate? What about the likelyhood of passing specific mental health legislation (like, firearm access restrictions) or the likelyhood of those restrictions having a measurable impact? Things like that.
Did you consider making these sorts of predictions and then decide against? If so, could you say why?
I don’t have much interest or knowledge in the politics of psychiatry like who becomes APA president. I am interested in the science of psychiatry, but I don’t think that changes in discrete ways on a year-to-year basis. Like I can’t predict “antipsychotics will fall out of favor this year”, because if antipsychotics were to fall out of favor it would be a multiyear very gradual process of decreasing use that was hard to notice. And I can’t predict “There will be a study showing antipsychotics are terrible this year” because there are studies showing EVERY drug is terrible AND great EVERY year, and it’s just a matter of synthesizing them in complex ways that don’t lend themselves to binary predictions.
Fair enough, thanks for the answer.
20. Hillary Clinton will win the Democratic nomination: 95%
Is it then implied that there is a 5% chance she will be charged with criminal misconduct wrt handling of classified material?
Or is this a combined Bernie/emails impact?
There’s probably a 1-2% chance of her dying or suffering health problems severe enough to preclude a run.
Good point, I shouldn’t have forgotten the age/health aspect.
According to this pdf (Chart 3) the odds of her dying in the next year given [female, age 68] are 1.4%, so I’d probably bump that up a hair to 3-5% for infirmity, even given SES.
Well, as regards “Oh Lord, she’s all of 68 years old, she could drop dead any moment from being ancient!”, the current president of my country ran for election back in 2011 at the age of 70 and has so far managed not to keel over in a deceased manner, even though people were making the same noises about him being so old and what about his health?
His election campaign was masterful; despite a public image of “dotty old grandpa and his amusing relic of socialist past days”, he outmanoeuvred them all – including the shark-minded ambition of those in his own party – and triumphantly sailed home.
Age and guile trump quite a lot of things 🙂
In the interests of keeping it clean, I won’t tell the young bull/old bull joke, but please rest assured I do take your point. 🙂
But that wasn’t what I was getting at – Scott assumes a 5% chance that HRC isn’t the nominee. I’m trying to tease out what impacts that 5%. To me, the 5% is much too low, given the variety of things which could be at play.
Looking at your comment just above this one, I’m tempted to ask “Isn’t predicting political psychopathology within Scott’s field of experience?”
Heh. Except that I’m not clear that’s quite what he’s doing, and if it is, it’s interesting how narrow his focus is. I’d expect him to branch out a bit beyond the front runners of the American election.
(And a question – why put this response here, and not under the comment above? I separated them out because I thought they were two separate things, and right-side limits on comments are a thing.)
Huh? What elective surgery?
I want predict that all your prediction on Trump will fail 🙂
As I side project I made a website for tracking predictions, https://www.yetipredict.com. It can help if you want to visual your predictions, compare them to other people’s, update them over time, or make ones which only you or your group can see.
As others have said, 18 (No country currently in Euro or EU announces plan to leave: 90%) seems optimistic, given the UK referendum.
Here’s my predictions sheet: http://oldlampsfornew.com/2016/01/07/predictions-for-2016/
Why so bearish on the Donald? And so ridiculously bullish on Rubio? He’s an awful, awful guy. And so bullish on Congress restricting gun ownership? The chance Congress will enact any meaningful gun legislation this year is less than 1%. Methinks you’re way too bullish on stocks and oil prices.
Gitmo will not be closed in the foreseeable future. I put the chance of it closing at .5%
To be totally honest, a lot of these predictions already seem less likely than they did when I wrote this a few weeks ago. I was also going to predict a peaceful resolution to the Malheur standoff (good thing I didn’t, I guess). It also seems less likely that anyone who supports Trump will be convinced to stop supporting Trump for any conceivable reason. And yeah, the market sucks.
If I were to redo it, the distribution would be very different — but that’s pretty instructive in its own way. I’m excited to see how it shakes out.
Not even caffeine?
Does anyone mind if I signal boost a project which I consider to have high value in science?
Ben Goldacre along with colleagues has recently set up COMPare.
For the last few years it’s been the norm that research in humans should be preregistered before the trial starts to avoid the file-drawer effect where negative trials don’t get published.
Without preregistration it’s hard to tell when someone has thrown a dart at a wall then built the dartboard around it.
It’s gradually been improving with a lot of research being preregistered but still published trials often don’t report what they said they were going to report or report things they didn’t preregister.
The dartboard is now there beforehand but people are still quietly building new dartboards around wherever the dart hits without mentioning it or mentioning that they were aiming at the original dartboard.
The COMPare project is doing something incredibly simple:
Reading the paper.
Reading the preregistered plan.
Posting a public note on their website and sending a letter to the publishing journal pointing it out.
It’s embarrassing for journals because in theory they should have made sure that the papers matched what was preregistered during peer review.
They’ve had a range of responses, some journal editors like the ones at the BMJ have posted corrections while others have doubled-down like the editors at Annals of Internal Medecine, it’s really quite entertaining.
You will dramatically change your mind on Psycho-dynamic and Cognitive Behavioral Therapy after reading/watching the following.
If you check any of this out (if your up to the challenge, especially the 5th link) you will thank me later.
What about the possibility of the North Korean government not surviving a war with another country? There’s even been a thinkpiece from a once-respected publication on it
If, hypothetically, you claimed to be able to calibrate predictions at 50.01% confidence, you’d require, on average, 34,657,359 independent predictions at this confidence level in order to provide me 1 bit of evidence in favor of this ability (assuming a simplified model where the only alternative hypothesis under consideration is that you actually have no predictive power on the predictions that you claim 50.01% confidence about).
(1 bit of evidence = the amount of evidence that would move a Bayesian from a prior “even odds” belief into a “2:1 odds” belief, or from a “2:1 odds” belief to a “4:1 odds” belief, etc. Yes, these turn out to be the same amount of evidence, no, that’s not obvious, but it’s okay.)
Here’s why: The great thing about measuring evidence in bits on a log-odds scale is that the value of each piece of independent evidence can be added.
So, how much evidence of your predictive power does one test of a 50.01% prediction give us? To make it easier, let’s give names to the hypotheses (remember, in this simplified model, these are the only two hypotheses to which we’re giving probability mass).
Hypothesis A: These predictions have independent 50% probability of being correct.
Hypothesis B: These predictions have independent 50.01% probability of being correct.
So, a “Correct” result is 50% likely under the hypothesis A, but 50.01% likely under hypothesis B. This is evidence for hypothesis B, in the amount of log_2 (50.01%/50%) = 0.000288510 bits
But an “Incorrect result” is evidence for hypothesis B, in the amount of log_2 (49.99%/50%) ~= -0.000288568 bits. (the negative indicates, of course, that this is really evidence against hypothesis B. But the math is easier if we embrace the negatives and call it all “for”)
So, if you’re really generating 50.01% correct results, each measurement will produce, on average,
(50.01%)(0.000288510 bits) + (49.99%)(-0.000288568 bits)
~= 0.00000002885 bits
You’ll therefore require (1 bit/0.00000002885 bits) ~= 34,657,359 of these observations to accumulate, on average, one bit of evidence in favor of hypothesis B.
Here’s a table for some other confidences you might claim, calculated in the same way, and in the format
Claimed confidence: Mean predictions required for 1 bit of evidence vs alternative “no knowledge” hypothesis.
Of course, there’s no entry for 50% – no observation will count as nonzero evidence of “you have 50% confidence” vs. “you have no knowledge” because, as has been pointed out at great length by other commenters, they’re the same hypothesis.
Generally I agree with this, but I think this only applies where their are only two discrete outcomes.
If we rolled a six-sided die, and you predicted the number 6 to come up with 50% confidence, that would be a true prediction.
Thank you, yes. The number of choices would affect our choice of alternative hypothesis; in your example, we’d probably want to judge my “50% probability of correctly predicting the die roll” against the alternative hypothesis “1/6 chance of correctly predicting.”
If you’re curious, each 6 rolled would be 1.585 bits of evidence for my predictive powers, while each other number rolled would be 0.737 bits against me. If I’m right, that information will accumulate at an average 0.424 bits per roll; if the alternative hypothesis is right, we’ll learn to distrust me at an average rate of 0.350 bits per roll.
Here are my predictions for 2016: http://economicsdetective.com/2016/01/predictions-for-2016/
I don’t know how to get into direct contact with you Scott, and can see you’re quite busy. We share some common interests and I occassionally want to direct some information your way. This one is relavent to a post you made some time ago regarding distinguishing religion from other areas of knowledge/endaevor, but doesn’t have a whole lot to do with the current post. Anyway, take a look at this interesting article: http://smithandfranklin.com/current-issues/On-not-defining-non-religion/9/16/117
Requesting a rule that one can’t say, “If you really believe X is likely, you can make a ton of money!” unless the poster also includes a link to whatever prediction market they are thinking of.
I didn’t include a link, but I assume that Scott can figure out how to buy bitcoins if he wants to.
Wow. There are a bunch of 90/95 predictions that I would put at 99+
Well, about #50:
Wait a minute. I thought Scott Alexander was self-describedly asexual. Now he seems to be a polyamorist. Is there an FAQ somewhere that explains this?
Personally, I love the ladies like Alpa Chino loves Lance Bass. But the idea of having more than one girlfriend at the same time … that just sounds exhausting.
IIRC he’s romantically attracted to women, he just doesn’t particularly enjoy boning them.
It’s entirely possible to have a sexual relationship without romance or being in love involved, and I think most people accept that. Now it is time to recognise that romantic relationships need not necessarily involve sexual activity, Greg. Catch up with modern times 🙂
(No, but it’s both funny and touching to me that at long last asexuality and aromanticism are being recognised. I’ve been Not Falling In Love as well as Not The Other Thing all my life and I appreciate the support about “Well, you don’t have to!” instead of “But won’t you be lonely unless you pair up as part of a couple?”)
“Never gonna fall for / Modern love” 😉
What Deiseach said – Scott’s a self-described asexual heteroromantic (see “Untitled”).
“47. California’s drought not officially declared over: 50%”
Here in Southern California, El Nino has been a bust so far and already we’re about halfway through the putative rainy season. Northern California, which is more important for the drought, has gotten more rain, fortunately, but the state has a whole lot of catching up to do.
I think we must have gotten your rain by mistake. It’s been pouring down for three months and is not really showing any signs of stopping.
Mebbe, when you caught that leprechaun, you should have been more careful what you wished for.
Ah now sir, one thing nobody in Ireland has ever wished for is “more rain” 🙂
No, but “more rainbows near me” is a proxy for the classic and often disallowed “more wishes.” The perpetual rain necessary to fulfill it is the unintended consequence needed for the fablular resolution.
47. I don’t manage to make it to my friend’s wedding in Ireland: 60%
Scott Adams says a considerable measure of peculiar stuff. I used to think he was simply screwing around and tossing thoughts at the divider to see what sticks until I saw the degree to which he truly guards his thoughts regarding sexual orientation legislative issues notwithstanding substantial feedback. I believe he’s only sort of insane. (Additionally an enormous misanthrope, however that is a civil argument for one more day.)
Autogenerated paraphrase of Merzbot’s comment above. Profile link is spam.
They can do that now? What was Scott’s prediction for robot apocalypse, again?
Thankfully, it’s little more than a lookup table of synonyms, and we’re some way from the horror of a marketing AI. No spambot apocalypse this year: 99%
(The auto-rewriting is known as “article spinning”, and it seems to have appeared in the spam/blackhat SEO arsenal around 2007)
Everyone is trumpeting the Donald these days but I’m surprised not to have seen anything about Bloomberg here.
Soda tax aside he was definitely a good mayor to live under, and unlike Giuliani he seems ‘presidential’ enough to have a shot. Being the mayor of New York is essentially the same as a governorship imo and if there has ever been a year for third party billionaires this seems like an appropriate one.
So what do people think about Bloomberg? Villain or menace?
>Being the mayor of New York is essentially the same as a governorship imo
You’re not up with the times, relevant administrative experience isn’t the name of the game anymore, it’s all about brand recognition.
Well then I’m sure he’ll get the coveted Bloomberg News endorsement at least.
I lived in NYC during his entire mayoralty and I’d vote for him in a heartbeat for almost any position, including President. I may not have agreed with all his crusades but he was effective, competent, and very much a net positive.
Third parties never win, only interesting question is who he’d be taking away the votes from. Pundits say it would be half-and-half, but I can’t imagine a Jewish New Yorker most famous for taxing things wouldn’t be drawing more Democrats than Republicans. If Trump got elected because Bloomberg split the liberal vote, that would be a really stupid way to end up in whatever crisis electing Trump causes.
Taxing things, promoting gun control, contraception and abortion, and the soft drink ban. So, yeah, not going to be a big draw for conservative voters outside of New York.
At the same time, things like stop and frisk, support for charter schools, fights with labor unions, and being a billionaire (especially one who made his money in finance) limit his appeal among liberal voters. He has an idiosyncratic collection of views that ensure that pretty much everybody can find something to hate.
Mmm. If one’s primary criterion for presidents is crisis avoidance, I think Trump has a great track record. The three colossal American mistakes of the last two decades can (I think) all be laid at the feet of Bush:
1. Bush opposed racial profiling of Muslims. (One of the airline clerks who checked in one of the hijackers was unnerved by him, but talked themselves down because “I’m just being racist, this guy is just a businessman.”) That same year, Trump called out Osama bin Laden as a threat by name (but overstates his prescience when recounting that story).
2. Bush went to war in Iraq, which Trump opposed (but he wasn’t public about his opposition until later).
3. Bush oversold the importance of home ownership and the creditworthiness of minority mortgage borrowers, leading to the financial crisis and Great Recession. I can’t imagine Trump making this mistake (and most of his real estate holdings discriminate on price in order to avoid the problems of low rent tenants).
Sanders wins points for his opposition to the Iraq War, but it looks like he either voted for or did not vote on the various mortgage acts. Clinton gets zero points. Cruz is too recent to have a record.
So looking forward I think there’s more reason to trust Trump to successfully navigate big issues than Clinton or Sanders or Cruz.
Maybe I’m just being cynical, but you can’t throw a rock in Washington without hitting someone who claims to have opposed the war in Iraq at its inception but who didn’t go public with that opposition until later.
Bloomberg might (10% < p < 30%), draw more Republicans than Democrats *in New York*. But that's not going to deny Trump a state he would otherwise have gotten. The only other places it might happen will be places where they think gun control is an effective method of crime control – states where the Republican vote is pretty low already.
@ Scott Alexander
Third parties never win, only interesting question is who he’d be taking away the votes from.
Sadly true. When someone looks like seriously splitting the vote of their own side, I question their motive and/or sanity. Sanders is old enough remember 2000, so I hope he stops in time.
Trump doesn’t seem that committed to either side, though.
I don’t have too much of an opinion about his politics (much of which makes sense for NYC but may not be sensible nationally), but the way he made his money is by building the communication and information infrastructure for financial traders. It’s recently come out that employees at Bloomberg probably accessed data on what information clients were looking at. So he makes me think “dystopian cyberpunk corporate overlord,” but I don’t have as much negative affect towards that as most people do.
He would probably be a better president than many, if not most of the other people running, but I estimate it as 90% likely he will have no appreciable effect on the outcome of this presidential race, whether or not he chooses to run.
I just realized that I kneecapped my own joke. That’s supposed to be “Bloomberg: threat or menace” up there a la the Daily Bugle.
If it makes you feel any better I saw it, realized what you meant, and appreciated the reference.
I am genuinely beginning to wonder if umpteen years of “Question authority!” being pumped into impressionable minds is having an effect, because yet again I’ve seen something on Tumblr. This time it was a repost of a Twitter about orange juice and apricot kernels cure cancer, because apricot kernels are full of “B17 and laetrile”, and yes, at least some people did comment that actually, guess what else is in apricot kernels? CYANIDE, and that this would be a great way to kill yourself (or a cancer sufferer) by following this advice. Also a disgruntled academic researcher was cross about the assumption that all they and their peers are doing is suppressing information for reasons.
But there were a fair few uncritical acceptance comments and at least one “Why don’t doctors want people to know this?” (which provoked the ire of the researcher).
As a conservative, I am grimly amused by the success of the notion that Da Man Is Always Puttin’ Us Down and so conventional [medicine, government, politics, etc.] is a vast conspiracy to keep people ignorant but there are these miraculous unconventional [cures, ways of ordering society] out there, when its success results in bilgewater like this.
But I still don’t want people to kill themselves and those they care about out of credulity. No, not even someone uncritically sharing T-shirt slogans (presumably the progressive version of bumper stickers?)
There’s probably also a lot of bleed-over from the Vitamin C cancer issue, which has gone back and forth since Pauling and started it’s own rather large industry.
For the interviews I just went to I ended up reading over a new report suggesting a plausible mechanism for why Vitamin C seems to be an effective treatment for certain colon cancers. I’m dropping a link to the study here but don’t feel obligated to read it: it’s interesting but you need to remember your glycolysis and pentose phosphate pathway units from cell bio to get much out of it. Point is that with huge concentrations of the stuff, which you could only get intravenously, selectively harms certain types of colon cancers.
But yeah, this sort of thing has been batted around in popular understanding even in my own very brief lifetime between “wash down your Vitamin C tablets with a tall glass of orange juice every morning” to “if you think grapefruit is good for you then you probably sharpen your knives with pyramids and collect crystals.” So I could see people latching onto some other vitamin / fruit as a cure-all and resenting scientific naysayers as a reaction.
I don’t think the problem is “Question authority!”.
If anything, science itself has slightly less questioning of authority than would be optimal. No, I think what you describe is more along the lines of being a massive idiot.
If these people weren’t disbelieving doctors telling them that cyanide is poison they’d be believing whatever “authority figure” had most recently published a self help book claiming that they can sure spider-bites with crystals thrice blessed by a druid under moonlight.
Often they’re doing that already.
Rejecting sane authority while embracing people like the awful poo lady (https://en.wikipedia.org/wiki/Gillian_McKeith) and believing anything the new authority figures say.
I predict (95% certainty) that I will be turning to the Hugo campaign for relief from the presidential campaign, and vice versa.
If you want something a bit more serious, any thoughts about who’s going to be running for vice president?
the Hugo campaign for relief from the presidential campaign,
Jimmie christmas, sing it, sista.
I predict at at least one point over the summer, it’s going to come down to getting off the internet longterm or getting a liver transplant.
Re: VP…I could have voted for Webb, depending on who the R was, so I’m open to him. My impression was that the D bench was still really shallow – is anyone else considered a strong candidate?
On the R side…Jeb won’t take it, but Walker might, and I’d be good with that. It depends on the primary candidate – if the prez is a strong outsider, then one would like to see an insider to balance the power inside the White House. And I’m not sure who of the insiders I’d like to have up there.
Assuming Hilary gets the Democrat nomination, wouldn’t it make sense for the Republicans to run a woman as VP?
It would be fun watching Hillary try to explain why the Republican candidate isn’t as womany as she is. Or even better, how the Republican candidate was a mere token and a distraction.
“They just picked her because she’s a woman, whereas I am running on my… Hey, did I tell you the story about how we got shot a… I mean, that whole Libya thing was just a ma… You have to admit, our relationship with the Russians has never been b… LOOK OUT! THERE’S A BEE!”
You can’t pick a woman just to pick a woman.
You have to start with some criteria in mind for what you need. If we think about the last 5 winning presidents (Reagan, Bush I, Clinton, Bush II, Obama), four of them picked Washington insiders because it was critical to their success in the White House (as they were coming in from the outside). 1 of them (Bush I) picked a fresh-faced Senator to balance out their old, grumpy, insider status.
The Republican candidates likely to take the nomination will each need a Washington old hand. Perhaps they won’t pick one because this is not that kind of year, but if they do, the stable of old Washington hand Republican women is thin.
Trump said a long time ago he wants Oprah to be his VP, and I suspect that’d be a solid play against Hillary.
Christ would also be a solid play against Hillary, but I see the odds of Him agreeing to run with Trump as only infinitesimally worse than the odds of Oprah agreeing to do so.
Trump should pick Ellen DeGeneres just to troll his own base.
Unrelated to the post, but a Google DeepMind program (AlphaGo) has beaten a three-times European Go champion, winning all 5 games of the match.
I don’t know much about the professional Go scene, but I’m not sure European players are the ones to beat. (Though it certainly is impressive.) What Dan is this player?
It was Fan Hui (2-Dan).
The point is that the curve has been so much faster than expected. It wasn’t that long ago that computers weren’t playing on full size boards and people were saying it would be decades at best.
The google program is going to play a 9 dan later in March (Lee Sedol). Even if it loses, the writing is on the wall.
They have a match scheduled in February against 이세돌, who is widely considered to be the best player alive.
Does your 60% Trump Nom and 60% Not Rubio Nom (i.e., 40% chance he will be nominee) imply 0% for Cruz and the rest of the field, or am I missing something?
Yes, but the predictions were made at different times and he has now since increased his non-Rubio chance to ~70%.
You’re wrong about Trump. If he wins the nomination (that’s the big IF) then he easily wins the general election. He’ll carry all the states that Romney carried plus NY and NJ.
It looks like it might be a Subway Series: everybody’s from NY. Trump is born and bred, Hillary is a carpetbagger former senator, Bernie is a Brooklyn-born Jew, and Bloomberg is the ex-mayor.
I posted my own predictions here: http://trotskyschildren.blogspot.com/2016/01/the-political-report-2016.html
Not as complete or imaginative as yours, but still.
Is it okay under SSC community norms if I OT/threadtangent this to ask a question about psychodynamic and cognitive-behavioral therapy? I’ve read S.A.’s article Scientific Freud (https://slatestarcodex.com/2013/09/19/scientific-freud/). The thing that I never understand about studies that compare different styles of psychotherapy is … do these treatments all cost roughly the same amount of money on average? Certainly, it’s impossible to imagine that anything called psychoanalysis could fail to break the bank (unless your analyst is a student, who is presumably studying so they can someday charge somebody else full price). But, if we are limiting the dicsussion to the relatively stripped-down psychodynamic therapy vs. CBT … and let’s assume that practitioners are charging the same fee per session … is the course of the average psychodynamic treatment going to require roughly the same number of sessions as CBT treatment?
If not, then any study which finds that the therapeutic results of two styles of treatment are about the same is actually showing that whichever one is quicker is in fact providing the same service for much less money, no?
Here’s the background of my question: for years, I have considered making a career change to become a psychotherapist. I am deeply attracted to the psychodynamic approach. Freud is the coolest. I have little interest in being a CBT practitioner. However, I remain deeply uncertain that psychodynamic therapists actually reliably provide a service to their clients at a reasonable price. I’m ambivalent about whether my own therapy has been very productive (I’ve certainly enjoyed it), even though I am (as evinced by the fact that I’ve considered a career in mental health) more psychologically-minded than average. Even if I’m convinced that most clients are deriving substantial benefit, if I’m charging them twice what they could be paying, that’s not a job I really want to do. It might end up being unsustainable to attract clients and, moreover, it’s just plain unethical if I’m aware that’s what I’m doing.
These concerns are a drag for me because they mean that I will probably not undertake the considerable challenge of changing careers to become a therapist, and therefore will most likely continue with my boring office job indefinitely.
P.S. to any of my office coworkers who somehow end up reading this: I love you guys; you’re the best. But surely you are also aware that our job is boring and contributes nothing to society.
We both seem to have a very strong interest in psychotherapy. There’s a new system of psychotherapy I think you will find interesting. The creator has made all the resources to practice it free, as any person who believes they have found a medical miracle would (the creator is Dr. Tim Carey)
I apologize for the reading list but this new therapy is based on scientifically sound theory. In fact it’s claimed to be the only therapy necessary to execute in the clinical setting. I’m more than happy to let you destroy these fantastical claims but before you do please consider the following.
First a quote from the book:
“I’ve become interested in different ideas. I have had experience or training in Applied Behavior Analysis; Familly Therapy; Cognitive Behavioral Therapy; Glasser’s Reality Therapy; Choice Theory (called Control Theory when I learned about it) and Quality Management; Ellis’s Rational Emotive Behavior Therapy; and Neurolinguistic Programming. That does not mean that I am eclectic.
I don’t know how to be eclectic. Each theoretical explanation implies a particular state of the world that is different from other theoretical explanations. How can you blend an idea that says “things are this way” with an idea that says “no they’re that way”? At a theoretical level I think of eclecticism as an impossibility.”
Now for the book itself: https://www.docdroid.net/Yml79Ks/da9lpedaypmc.pdf.html
If you have read a bit and find it interesting more free resources our here:
I have been trying to show slatestarcodex but so far he hasn’t noticed.
So Scott A (the one who writes this blog, not one of the other two) is still a “resident”. That means he has to perform his psychotherapy in the Approved Manner, and is subject to questioning by his nominal supervisor. So even assuming that a) your system is better, and b) Scott is interested in it, he can’t look at it now.
He has a BIG hoop to jump through in the next couple of years before he’s allowed out on his own, and if he’s seen as being influenced by your system before he’s done with his residency, there’s a big chance he’ll be judged as having missed that last hoop jump.
Once he’s licensed and has a practice, he can consider alternate theories of pyschotherapy. Right now, he can’t.
In order to properly calibrate yourself, it’s even better to answer questions posed by others! We should therefore all join in…
I’m posting (most of these) on Predictionbook until I get tired of doing it; the first 50 of these (not the personal ones,) plus perhaps the ones form the other blogs linked. Feel free to add more, and mention it here so we can all participate!
See what I’ve been posting/predicting:
Also, GJOpen seems like a nice place to make predictions on events others have picked, and get better calibrated; https://www.gjopen.com/
This is ludicrous, considering that FiveThirtyEight gives her a 76% (at highest) chance of winning Iowa and an 18% chance of winning New Hampshire. Yes, there are multiple ways to arrange states to get the nomination, so the Sum Law plays in, but any nomination win involves an arrangement of winning more than one state, which makes any given arrangement subject to the Product Law instead. Since the various arrangements have strong temporal and causal dependencies (ie: losing New Hampshire crosses out any victories that involved New Hampshire’s delegates), I don’t see how you can get a 95% chance of nomination out of any combination of states where the individual probabilities are a good deal less than 95%.
I could just be bad at math, but I think in this instance I’m not.
If all 50 states had the same weight, and if someone was 76% to win every one of the 50 states, then assuming independence (which is obviously a huge assumption and does push the number upwards), they would have above a 99.99% chance of winning the election. You’d only need about a 60% chance of winning each state to get a 95% chance of winning the election.
> North Korea’s government will survive the year without large civil war/revolt: 95%
I estimate 93%
You do realize that you’ve implicitly assigned probability 1 to “Rubio or Trump wins the republican nomination”?…at least conditional on a single individual winning the republican nomination and that Rubio and Trump are distinct individuals.
> 3. Greece will not announce it’s leaving the Euro: 95%
I say 93%