20% chance you will have a child, 11 months out, means either you are making a math error or you are trying to conceive. Assuming it's the second, so good luck!
I think ‘no reasonable person would invest in this’ is an incredibly low bar. Surely lots of intelligent people with money to spare make the occasional small moonshot.
Exciting that we can compete alongside your predictions this year!
A small suggestion: perhaps move away from using the Dow in future? It is a nonsense index with wacky methodology that made sense 100 years ago, and is mostly used by the media for historical reasons/financial illiteracy. S&P 500 (SPY) or total return (SP500TR) are much better.
"supreme court will overturn roe v wade" - this may be difficult to adjudicate without clarification. By some standards it was already overturned (read: "clarified") by Casey. I suspect you mean, and it might be helpful to rephrase it to, "the supreme court asserts there is no constitutional right to an abortion".
Things are looking very dire for Bolsonaro around here right now. Most polls place him with around 25% of votes in the first round while Lula has around 45%. And the hypothetical rerun polls put Lula with a ~20% advantage too.
The economic recovery - which was Bolsonaro's big bet for a rebound on the polls - doesn't seem likely. Inflation is still above 10% and the increase of interest rates will most likely slow the recovery. Unemployment will probably remain very high (around 12%).
Bolsonaro's entourage was also betting that we would have overcome the Covid crisis and it wouldn't hurt the president's popularity as much anymore. However we are seeing that even with a much lower death and hospitalization rate, his approval won't go up. People are not forgetting his mismanagement (and straight brutality) during the pandemic that easily.
All in all, I would lower that 50% to about 20 or 25%. Bolsonaro will probably not get reelected.
I take it that second-guessing the Breyer retirement seems too easy now, but I’m thinking there is still room for a prediction. He has said it is contingent on getting a successor, and it’s not completely clear to me that this will happen.
Hi, I'm the co-founder of Manifold Markets (with 30% chance of lasting through 2022 :P).
We've created prediction markets for all of Scott's predictions that were included in the contest!
Currently, you get 1000 Manifold play dollars if you sign up that you can trade with! It's fun, and the best traders will be tracked in the leaderboards for Scott's predictions.
How are we measuring prediction market size? I'm looking at alexa siterank, but that's only ordinal rather than cardinal so it doesn't make sense to talk of things being half as popular if that's the resolution source. Most of them don't publish anything about daily active users, total money transacted, etc, so the only metric I can find about all of them is alexa.
I entered the contest. This is the full list of all our large disagreements (defined as one of us thinking either the YES or the NO side is >1.5x more likely than what the other thinks.)
Biden approval rating (as per 538) is greater than fifty percent (40)
25
PredictIt thinks Joe Biden is most likely 2024 Dem nominee (80)
90
PredictIt thinks Donald Trump is most likely 2024 GOP nominee (60)
75
Major flare-up (worse than past 10 years) in Israel/Palestine conflict (5)
8
Dow above 35K (90)
80
Inflation for the year below five percent (90)
75
Unemployment below five percent in December (50)
80
Google widely allows remote work, no questions asked (50)
35
Fewer than 10K daily average official COVID cases in US in December 2022 (20)
5
India's official case count is higher than US (5)
20
Some new variant not currently known is greater than 25% of cases (60)
75
Most people Scott sees in his local grocery store on December 31st are wearing masks (60)
90
Masks still required on US domestic flights (60)
90
China has fewer than 100,000 COVID cases this year (official estimate) (30)
70
New legal US real-money prediction market at least half as big as Kalshi (5)
10
New illegal but easy-to-use market satisfying the above (20)
30
Inflation for the year under three percent (FP: 80; SA: 80)
40
At least one country will have less than 10% of people vaccinated with two shots by November 1st (FP: 70; SA: 95)
99
A psychedelic drug will be decriminalized/legalized in at least one more US state (FP: 75; SA: 75)
50
The Biden administration will set the social cost of carbon at $100/ton or more (FP: 70; SA: 70)
35
Joe Biden is still president (MY: 90; SA: 90)
96
At least one Biden cabinet-rank official resigns (MY: 70; SA: 70)
>35. No current residents leave our housing cluster: 80%
Does this mean "nobody moves away" or "nobody moves away AND nobody dies"?
>9. No military conflict between the PRC and Taiwan (a worryingly low 90%) HOLD
My prediction last year (not sure if I posted it on here) was "30% nuclear war over the Far East within the 2020s". I'm not plugged into current events well enough to give a year-by-year. The implicit few% nuclear war this year seems fairly consistent, though.
Could we have the contest at the top next time so that we can give the predictions a shot before looking at your answers? As someone with no experience forecasting and knows that you have historically been well-calibrated, I feel like I should just copy your answers wholesale.
“ 95% chance they lose both houses of Congress implies 97.5% chance of losing each house”
I don’t think that’s necessarily true, seeing as P(democrats Lose the House) and P(Democrats lose the Senate) are not independent. It seems like you could say P(Dems lose House) = 0.96 and P(Dems lose Senate I dems lose House) = 0.99 so P(dems lose Senate)*P(dems lose house) = 0.96*0.99 ≈ 0.95
Now that I’m writing these out, I agree that 0.95 is still to high, but its possible to have the conjunction of the two be 0.95 while they could individually be any number such that 0.95 < p(
4. Supreme Court will overturn Roe v. Wade (65%): SELL to 60%
There must be something I'm missing. This % seems really high given the base rate of overturning previous decisions is so low. I would have thought a straightforward overturning of abortion would ~30% likelihood with the more likely scenario(80%) they cripple roe vs wade.
Eternal glory is fine with me - but who wants "a free ACX subscription"? The point of subscribing ACX is the wish to show Scott how we really, really appreciate (I do) his blog by doling out some dough ( I did not. Yet.).
I tried to skew all my answers to the contest based on a few specific guesses about the future, rather than trying to be accurate. This should hurt my score in expectation, but won't it increase my overall chances of winning the contest by skewing my score distribution downward?
* I don't think I've ever been *explicitly* called a superforecaster, but SciCast once wanted to interview me because I was on their leaderboard. Does that count?
* What do you mean by "within two generations" exactly? Would a second cousin count?
"Biden approval rating (as per 538) is greater than fifty percent: 40%"
This strikes me as *very* unlikely. While better Covid situation might improve things, I wouldn't give this a better than 10% chance. *Obama* didn't have this after two years!
"Inflation for the year below five percent: 90%"
You should probably specify exactly which inflation measurement you're using here.
"China has fewer than 100,000 COVID cases this year (official estimate): 30%"
Huh. So *fewer* (even at just 30%) cases than in 2021, with Olympics and Omicron putting pressure on the situation? Well, I guess we're talking the official numbers here, so who knows?
What exactly do you predict here? I could imagine at least three reasonable interpretations.
a) Starship is at an orbital height with velocity no less than orbital velocity at that height, and with direction of travel closer to horisontal than vertical, up vertically, and also very close to the pre-announced plan. This covers «clearly could go orbital but deliberately opted not to, presumably for failure impact mitigation».
b) Starship's position and movement are clearly orbital at some instant, even if an orbit is never completed.
c) Starship either completes a full orbit around the Earth above 100 km line, or intentionally releases an active payload that does.
> 29. Most people I see in the local grocery store 12/31/22 are wearing masks: 60%
Does this come with an implied «I go to the grocery store on 2022-12-31: >95%»? Or is it «last time in December 2022»?
> V10. China will not reopen its borders in the first half of 2022 (80%): BUY to 90%
What level of reopening? Business travel for non-state-level reasons in at least one direction, possibly with quarantine on top of quadra-shot and testing, or tourism in both directions with at most 7-day quarantine (with easily available level for US and China residents levels of vaccination and testing)?
> V17. AI will discover a new drug promising enough for clinical trials (85%): HOLD
For what values of discovery? Is something like https://www.science.org/content/blog-post/alphafold-excitement working out enough? (Human specified targets, specified search space, significant help from AI in filtering, estimated time savings in single-digit weeks)
> Y25. Russia does not invade Ukraine (60%) HOLD
I think it is a good idea to precommit to an information source choice procedure for the resolution. given that there might be pairwise difference between claims by Ukraine, Russia, and US. (It is also possible the difference won't be relevant for resolution).
Also, what level of sabotage/riot incitement infiltration as estimated by the chosen sources would count as invasion?
"9. Honduran ZEDEs legally crippled to the point where no reasonable person would invest in them further: 5%
10. New ZEDE approved in Honduras: 30%"
Given that I think no reasonable person should invest in Honduran ZEDEs anyway, this is going to be splenetic on my part.
So I think it depends how corrupt Honduras/Honduran government is, and the cursory Wikipedia search shows that it's pretty corrupt. After all, this is why ZEDEs picked Honduras in the first place; a poor country where the government would be happy to look the other way about setting up private fiefdoms so long as the right money changed hands "for the economic development of the area", naturally.
So if Honduras is *moderately* corrupt, I think that yes, there's a low chance on no. 9. If it's *very* corrupt (and the new guy seems every bit as scandal-ridden as his predecessors and is a member of the same party which has such lovely items on its record as "In May 2015, Radio Globo discovered documents that allegedly showed that the Honduran National Party had received large amounts of cash from nonexistent companies through fraudulent contracts awarded by the IHSS when it was run by Mario Zelaya. The contracts were approved by the National Congress of Honduras when Hernández was its president and the party funding committee was headed by his sister, Hilda Hernández. Hernández has accepted that his election campaign received money from companies tied to the scandal, but denies any personal knowledge. By June 2015, Hernández had appointed a commission to investigate the cause of the corruption." It's very convenient when you are able to appoint the commission that is investigating your own alleged corruption!), if it is *very* corrupt, then the chances go up.
Because ZEDEs are cash-cows for the governments of the countries involved, and what good is a cow you can't milk? The ZEDE for market gardening/horticulture should do okay, South American countries are familiar with the model of having peons to work on plantations for the benefit and enrichment of a handful of eminent and prominent families.
The ZEDEs which are "entice rich Westerners to come here and work remotely at their high-paying jobs while enjoying low cost of living and dirt-cheap local labour which will have no rights under local law since *you* will be the local law, hence every stakeholder can be lord of the manor" are going to be milked *hard* because you're not producing anything locally like the giant glasshouses/polytunnels ZEDE, and that will mean all kinds of levies, taxes, charges and other "Hello, I am the Minister for Brown Envelopes, so long as you keep shovelling money into my private Swiss bank account your holiday resort will be left alone". No, I can't put a number on this. Yes, I think this bubble will eventually burst.
"4. Supreme Court will overturn Roe v. Wade (65%): SELL to 60%"
They're giving as high as 65% for this? Well, given that this is a perpetual scarecrow for liberals ('they are going to go after abortion rights!!!!' every time there's a Supreme Court nomination coming up), I'm not that surprised that Vox are getting ants in their pants about this, with the Breyer announcement and a majority Catholic court (and believe me I am immensely tickled about the spectre of the Horrid Popish Plot being dragged out of its grave once more).
I don't think it's going to happen. Things have moved more on rolling back abortion than I ever thought would happen, see the Texas laws, but there was swift movement to combat this, see the Texas laws. So I'd put it more around 40%, if even that high. A possibility because it's always a possibility, but nowhere near a likelihood.
"5. Stephen Breyer will retire from the Supreme Court (55%)"
"21. Kenneth Branagh's Belfast will win Best Picture (55%)"
Ooooh. This one. I was only very vaguely aware he had a new movie out, and looking at the synopsis. Well. Yeah. Standard Troubles movie. I don't know if it's strong enough to win, or what movies it is going up against. I would say 'no' for various reasons (e.g. it's too white, not diverse enough, for modern tastes) but maybe it will be seen as a safe bet to vote for because it's not American?
The real problem is that this year is the 50th anniversary of Bloody Sunday, something which is still very raw and still affecting relationships between Ireland and Britain, and "Belfast" seems a bit too *safe* a treatment of the whole tangled mess:
For a start, this is Branagh reflecting on, or presenting a version of, his childhood in Belfast. Bloody Sunday was in Derry. It's set at the start of the real breaking out of violence, before the British Army becomes the villain of the piece. It's on a personal level of a family and a small boy. It's extremists on both sides who are the problems, with a sort of "why can't we all just get along?" message. It may appeal to voters as a safe choice that deals with Issues that are in the past (the 60s) and far away (Ireland) and so they can have their cake and eat it.
I don't think it will win, though. I think it'll get some awards but not Best Picture.
While we're talking about the North:
"27. Sinn Fein becomes the largest party in the Northern Ireland assembly (60%)"
At the moment, the two major parties are the Democratic Unionist Party with 27 seats and Sinn Féin with 27 seats, the remainder made up of odds and ends:
It looks like they'll pick up some extra votes, but the thing is that the parties cannabalise each other; for Sinn Féin to get extra votes, they will peel those away from SDLP and smaller parties. The DUP will do the same by absorbing Unionist votes and replacing, say, the 1 seat held by a small party candidate. As you can see from this news story, where the 1% for Sinn Féin has come from the SDLP losing 1%:
"The gap has widened since the last LucidTalk poll in November, with Sinn Féin up one point and the DUP down one.
Of the other main Stormont parties, the poll puts Alliance and the Ulster Unionists in joint third place on 14% (Alliance down one point on November and no change for the UUP), the TUV on 12% (up one) and the SDLP on 11% (down one)."
Can you compete if you already looked at prediction markets before you posted, and have seen many of the questions, but don’t look while making your submission?
If you'd like to play D&D, I'd love to run the game for you and a friend. I've been DMing a while, and I've been interested in your writing since you were posting on LW as yvain. A lot of the themes in your work (particularly The Goddess of Everything Else) touched me deeply and affected my own writing (I wrote Significant Digits, the HPMOR sequel). Happy to do a one-shot trial first to see if the vibe works.
60% chance of Donald Trump being the GOP candidate? So that means you think there's a 40% chance he will die in the next 2 years? Because that's the only way I see him failing to get the nomination, such is his grip on the party.
(28) seems like it might need a little more clarification? Does a "new variant not currently known" need to be one assigned a new Greek letter, or have some number of mutations relative to current variants? The media's been very excited lately over possible vaguely-variants like "delta-cron" or "omicron 2" or whatever, so I can easily see a future where something the media calls "mega-delta-omi-doomy-zord" or something is responsible for a large number of cases, even though it isn't technically a distinct variant by the standards used by official health agencies.
The odds for democrats holding one of two houses seems way too low. Currently betting markets only have republicans with 85% chance to win the house and 75% chance to win Senate. So 90% chance they win both seems overconfident.
The temperature forecast also looks overconfident towards warmth. In the last 30 years only 60% of years have been warmer than the year before them. Even in the small sample of the last 10 years its only happened 70% of the time.
1. SELL to 20%. Biden is currently at 42% (as per 538). Getting to 50% would mean improve his polls by +8 pp. This is better than any president has done between the end of their 1st and 2nd year since Truman. Most presidents' approval ratings decline during this time period.
9. BUY to 15%. The recently elected president explicitly campaigned against them. The ZEDE law is in the Constitution, which should make it hard to overturn, but not nearly as hard as it would be to change the US Constitution. How would this resolve if the government passes (and enforces) a law which blatantly contradicts the Constitution?
ECON/TECH
19. BUY to 60%. Unemployment is currently 4%. So you're thinking that the unemployment rate is more likely to go up. 4% is unusually low, but the unemployment rate usually trends downward, and only occasionally rises rapidly. I also don't see how this is consistent with Biden's approval rating increasing +8 pp 40% of the time.
COVID
You have predictions on how high the floor will be for endemic covid, but not for how high the ceiling is. This seems like an easier and less useful prediction.
25. BUY to 10%. So far in the pandemic, India's case numbers have been higher than the US for slightly more than 2 months (May & June 2021). This is about 10% of the pandemic. The US also has a smaller population, is better vaccinated, and probably has had a higher percent of infections.
28. BUY to 80%. This has happened at least 2 times per year so far in the pandemic. I also think that this is more likely to happen than WHO designating it as a variant of concern (VOX 13) - especially since the new variant would have to be more infectious, but not necessarily more dangerous.
29. I'm not familiar with where you live, but I hope this isn't true. I suspect that this will be true for most of the country by the end of March. Which is a long time before December.
32. SELL to 20%. This will only get harder with time. New variants will be more infectious. There will be more pressure to open borders as the rest of the world returns to closer to normal.
BLOG
77-79. It's interesting to get an order of magnitude estimate of how much this blog makes.
82. BUY to 70%. Even higher if different cities in the Bay area count as more than 1.
87. This is a "Science" blog? There isn't even a Molecular Monday for new drug tests.
VOX
3. Don't you have a 50% chance that unemployment will go up to 5%?
10. This is necessary if they're going to keep a zero covid strategy.
22. Sell to 40%. Norway is probably the favorite, but I don't think that any country should be above 50%.
YGLESIAS
22-23. Inflation is often a self-fulfilling prophesy. If people think it will happen, it will, and if people think it won't, it won't. Vox is much higher profile than Yglesias. Vox doesn't want there to be high inflation, and so might purposefully underestimate it, in the hopes that they can influence it. Yglesias is low enough profile that he probably doesn't think that his prediction will impact what happens. I would lean towards Yglesias's prediction.
Assuming I wanted to duplicate Scott's bets because I revere him as an insightful man of foresight and wisdom, where can I do that? I'm moderately familiar with the concept of prediction markets, but I've never placed a bet.
> 18. Inflation for the year below five percent: 90%
(under Vox)
> 2. Inflation in the US will average under three percent (80%): HOLD
(under Yglesias)
> 22. November 2022 year-on-year CPI growth is below 6% (70%) BUY to 80%
> 23. November 2022 year-on-year CPI growth is above 4% (70%) SELL to 50%
Isn't CPI growth the same thing as inflation? Or at least, it's one of the standard measures of inflation. It seems odd to have inflation under 5% but CPI above 6%, and yet you're assigning at least a 10% chance of that. Similarly, it seems odd to have inflation under 3% but CPI growth above 4%, and you're assigning at least a 30% chance of that.
I have no real comment on the predictions, but I have to laugh darkly at the phrasing "at least $250 million in damage from a single round of mass protests".
Traditionally, one would not refer to events causing upward of $250 million in damage as "protests".
I don't know what "loses power" means here precisely, but a 60% estimate of him losing the election is wildly off. In my experience, basically every sensible person in Hungary thinks that a small-margin Orbán win is the most likely scenario. I mean, it's not impossible that he loses, but it's surely below 50%.
As a brazilian , I can´t agree with this bullish analysis of Bolsonaro´s chances of reelection , 50 % seems a lot right now. He is pooling worse than any other who was able to be re-elected (i.e: Temer does not count). I would to read counterarguments
"at least $250 million in damage from a single round of mass protests"
How is "single round" defined?
Can that be n days in m cities with n, m > 1, assuming they all those individual protests share the same cause (assuming there is a specific one, that can be pointed to)?
Yglesias is too optimistic about inflation. My predictions for "Consumer Price Index for All Urban Consumers" change over 2022:
≥5% increase 90%
≥6% increase 70%
≥7% increase 50%
≥10% increase 20%
This inflation is caused by Trump and Biden printing money to pay for the Covid stimulus at the same time that overall production fell, and it's going to take at least a couple of years to work through the system.
Surprisingly, in my 5 minutes of research I adjusted way down on Yglesias 2 and a little up on Yglesias 3. I was expecting to agree with 80% since it's likely to be a red year and the Senate skews red anyway, but the particular seats up for election look good for Democrats. They aren't defending any Trump-won seats, and I expect these midterms to lean red rather than be a red wave. That means the states I judge to be at risk of flipping are Georgia, Arizona, Colorado, New Hampshire, and Nevada. I didn't have time to look up each individual approval rating, but if I remembered right, most of these Senators are reasonably popular in their states. I think Georgia is more likely than not to flip, but the rest are (individually) more likely than not to remain blue. Altogether I do think at least one will flip, but I wouldn't put my confidence at 80%.
As for Yglesias 3, the next-reddest states after those 5 are Illinois and Washington. Even in a wave election I don't see those flipping. 80% is too low.
As long as people are hanging around here forecasting, might as well enter the 7th annual version of my contest, which looks much like this: https://forms.gle/P3tR8xccsYjYTi5A9 (late entries are fine, just not eligible for the prizes: a book and a bowl of pho)
how is >66% US population fully vaccinated (by current standards) against COVID being evaluated? does this mean Jan 2022 standards (not boosted), Jan 2022 standards (boosted), or the standards in Dec 2022?
I like how Yglasias specified CPI when talking about 'inflation'. i know there is a general understanding that when officaldom talks about inflation they mean the CPI, but they are truly not the same thing.
I find many econ ideas are purposefully or needlessly convoluted and looking into them to unpack what they are saying means a lot. Out of the three big ones I hear most often, GDP, Unemployment, and Inflation....the offical numbers are representing a very narrow and very politically driven set of figures which get reported.
I'd encourage anyone to research further into what is and is not included in the CPI number and see if you think that value represents the inflation you experience in your own life.
It does not include any information about the money supply or real estate prices or industry. It also uses a crazy substitution principle to try to 'mimic' what real world buyers might do.
So if the price of beef goes up and chicken is cheaper, they'll assume for you that you'll switch to buying chicken instead. So.....this number is NOT telling you about the changes in the price of goods then. It takes real inflation and a real upward movement in prices and finds a way through biased arbitrary selection to remove that inflation from their calculation of inflation.
Same with electronics which are one of the worst offenders. Buying a new phone or computer after 2-3 years? Say you're a 'I want the flagship phone' type of buyer...If that phone was $1,500 a few years ago and is $2,000 today, then the price you'll pay has gone up for 'the same' product of the flagship phone.
But they'll pretend that the new 2nd or 3rd tier phones are equally if not more powerful than the ones from a few years ago or have longer battery life or any number of arbitrary selections they make to pretend you can now get 'the same' phone for $800 instead. Every single year for around 20 years they've had this nice negative figure to modulate the overall basket in any way they wish using whatever choices they want to make that year in officialdom.
If you have this much freedom of choice or degrees of freedom in what you're doing...then you can say almost anything you want to say. In a deep dive on this, even 10% of this level of freedom of choice would see Scott lambasting a psychiatric drug study and saying how we'd be utter fools for using it or relying on it to tell us about that drug. They can pick and choose every single variable and arbitrarily include or exclude things such that every year the CPI 'metric' is measuring a different thing while attempting to lie to you about how they're measuring the same 'conceptual' idea of their guesstimate of consumer choices...a concept whose components are redefined every time.
And do we go back and update the CPI to a 'real one' after we interview lots and lots of people to get a refined measurement of consumer behavioural changes in buying choices? Nope! There is not even a hint of a notion that officaldom would ever do that with their CPI.
They are not doing science or the lower religious artform of economics, they are not seekers of truth or trying to tell us something worth knowing about the world. The lack of updates, accuracy, or even an attempt to do this tells you all you need to know. The official CPI is a political and propaganda tool to influence financial markets first and foremost.
It is a pretend game of counting and not counting major categories of prices which affect people's lives...real estate up 30% in your area while Blackrock goes on a buying spree spending 20-50% above market rates? Buying a home is the single largest purchase in the average person's lifetime....don't worry! We DIDN'T count that in 'inflation' or CPI specifically.
Truly this is inconceivable! In the classic meme sense of Princess Bride 'I do not think that word means what you think it means'
CPI is so wonky, manipulated, politically driven, and diverts so very very far from reality and what people's lives are like that I'd assert it cannot be anything other than a lie and is a part of the propaganda machine more than it is is part of any academic or trustworthy way to attempt to describe changes in the real economic landscape.
against FP you have Emmanuel Macron at 65% but against MY you only have 60% (technically I could see it the other way because MY only says reelected but FP has reelected president of France so it should be X% and X.0000001% where EM is appointed to some other position and then runs and wins reelection)
I would like to see some more skin in the game predicting about education around here. Too much Bryan Caplan and Freddie DeBoer (sorry guys), plus motivated reasoning inform the discussion. As a first step in a remedy to this here is a market that is OBVIOUSLY mispriced. https://manifold.markets/JohnBuridan/will-any-us-state-excel-its-previou
Fun! for the betting market, I'm just following Scott.
Except for the chance of kid, If he's really trying to make a baby, I'm saying ~80%, but I don't know that's true so I'm going to go 50% of making a baby... which is betting with my heart. :^)
Having a kid changes your whole outlook on life. It's magical and a pita.
maybe off topic, but not sure a better place to answer a q i've been exploring a while: what is the best corporate prediction market software out there? looking for something cloud-based to build out with teams, and my hacked version in google sheets doesn't exactly inspire confidence or confidentiality.
Very Interesting. I also made the same predictions for the contest.
There are a few areas where I disagree to a major extent with the predictions offered by Scott Alexander, Vox Future Perfect and Matt Yglesias. Here is a quick rundown (those predictions with more than 25% difference)…My predictions are in the row below the original prediction.
- PredictIt thinks Joe Biden is most likely 2024 Dem nominee (80)
My prediction: 40% - I feel like if Democrats loose the midterms, then Biden will just resign on his own and Harris will step up. Also, other Democrats could be making a bit of a fuss about him if he doesn't deliver...
- Dow above 35K (90)
50%: Already, the Dow is showing signs of struggle. If the Fed raises rates, then it's hard to see the Dow to continue to being as strong...
- Inflation for the year below five percent (90)
50%: That was already said last year that inflation would only be "transitory"...well that wasn't the case ....so why should it change this year...only if the Fed raises the rate quite a bit...but will they?
- Starship reaches orbit (90)
50%: To be honest, I don't know much about it, but considering Musk's track record in getting products to the markets in a timely fashion (Roadster?), I don't think it's that realistic...
->66% US population fully vaccinated (by current standards) against COVID (70)
95%: Didn't that already happen? Anyway, looks to be veeery achievable …
- Masks still required on US domestic flights (60)
30%: I don't really see the point in this...if there are tests before every flight, and/or vaccine mandates, then why insist on masks (especially if they are just surgical masks?).
- New legal US real-money prediction market at least half as big as Kalshi (5)
50%
and -New illegal but easy-to-use market satisfying the above (20)
50%: I don't know much about these, so I just said 50%...
- Inflation for the year under three percent (FP: 80; SA: 80)
10%: That seems like a ridiculously high number...I think it's really unrealistic to have substantially lower inflation this year...oil prices are still rising, as are many other input prices for many goods...
-Democrats will lose their majorities in the House and Senate (FP: 95; SA: 90; Matt Yglesias also predicted this event, saying 90)
50%: Yeah, like I said this seems to bullish on the Republicans...they probably will win the house, but both the house and senate? That didn't happen in 2010, and while Biden is now less popular than Obama was back then, the difference isn't that huge, and Republicans are now much more of a non-mainstream party...
-Supreme Court will overturn Roe v. Wade (FP: 65; SA: 60)
40%: This also seems rather high...I mean it's been the law of the land for so long...don't think most judges would want to change it, and opening up a whole new can of worms...
- Jair Bolsonaro will be reelected president of Brazil (FP: 55; SA: 50)
30%: Well, he's badly trailing in the polls, and seems to be very unpopular...of course they said that about Trump, but Lula is no Hillary...so except if he rigs the election, or Lula gets jailed, I don't see him winning...
- China will NOT reopen its borders at any point during the first half of 2022 (FP: 80; SA: 90)
60%: I am not sure...how long can they really go with this strategy? Probably this year, but then they would have to rethink if it's not a net disbenefit...
- 20% of US kids between 0.5 and 5 years old will get at least one COVID vaccine by year's end (FP: 65; SA: 65)
40%: Don't think it's that realistic to be honest, as so many anti-vaxxers are people with children...also, maybe there won't be as much pressure for people to get vaccinated anymore?
- A psychedelic drug will be decriminalized/legalized in at least one more US state (FP: 75; SA: 75)
50%: Hard to say, but I would be less bullish on it...Cannabis yes, but these types of drugs (LSD?) seem not to be as popular with the broader public so IDK...
- AI will discover a new drug promising enough for clinical trials (FP: 85; SA: 85)
50%: I don't think AI is advance enough....
- Norway will win the most medals at the 2022 Winter Olympics (FP: 60; SA: 60)
30%: Not sure why a small country like Norway should be the favorite? I mean Russia, the US or even Canada sound more like favorites to me...
- Democrats lose at least two Senate seats (MY: 80; SA: 80)
30%: Which ones? I think they might loose NH, maybe GA but probably not AZ... Though I think Democrats can win PA, and possibly WI and/or NC...so on net its -1 to +1, so 0 net seats on average...
- Liz Cheney loses primary (MY: 80; SA: 80)
50%: I don't know...possibly, but I would give her higher chances than this though...
The predictions from Macron seem ridiculously low. I've seen an analysis from the Economist (https://www.economist.com/interactives/france-2022) that gave Macron 90% odds of winning, and that matches both my intuition and the polls. His only remotely serious opponent is Pecresse, and he's still projected to win against her (I'm really puzzled anyone could prefer Pecresse over Macron, but I could see Pecresse winning). Everyone else has no chance.
I'm not into prediction markets, but I'm tempted to actually get in on this one. Either the market is seeing something I'm not or it's suffering from american-centrism and guessing wildly off the mark.
Inflation <5% at 90%. Really? I would not dare to put that higher than 50%. It is already spiking at 7% right now. Labor shortages, oil market is very tight with limited drilling inventories in US shale fields at lower break even prices. And supply chains will see more pain with China sticking to zero Covid in 2022.
If you put a major flare up in Ukraine at 50%, then that alone could push inflation above 5%. Since Russia is a major commodity supplier. And any interuptions to export could send commodity prices flying.
Nobody seems to have mentioned this: at the beginning of 2018 (slatestarcodex.com/2018/02/15/five-more-years/) you gave Roe v. Wade being overturned a 1% chance. Now you give it 60%. Regardless of what actually happens, at this point you'd have reduced your money to a 60th of the total if you'd bet on this.
I'm amazed anyone would ever give Roe being overturned a 1% chance, let alone under a Republican president and senate, with 2 or 3 liberal or moderate justices soon likely to die or retire, let alone someone as good at prediction as you. I guess I'm just asking (light-heartedly, with no offence intended): what on earth were you drinking when you made that earlier prediction?
A bullish position given that the Dow has been hovering around 35K recently. Why so much confidence?
> No recession in 2021 (90%) SELL to 80%
Hmm, what kind of bull is this?
> Medical establishment reverses course and officially says any of Vitamin D, HCQ, or ivermectin is actually effective against COVID: 1%
Also surprising. Earlier you gave a 10% chance that ivermectin works ... is there new data that made you update even more against ivermectin? Or is it just that you think they won't "officially" support it unless the effect size is shown to be unexpectedly large?
> Fewer than 10K daily average official COVID cases in US in December 2022: 20%
...for reference this would be below the 2021 low point on June 20, 2021, and below the low point during the 2020 lockdowns. So, it's obviously unlikely: it would require no new variant-of-concern to arise, and it would also require immunity that endures long enough to prevent flare-ups. Alternately, a genuine apocolypse could cause this result, or a halt to the gathering of statistics.
> At least three dates with a new person: 30%
This I did not expect, Mr. Groom, especially since you said in 2014 that you were asexual.
> The Biden administration will set the social cost of carbon at $100/ton or more (70%): HOLD
Wait, what effect does this number have? I thought it seemed very high as a carbon tax, but low as a estimated social cost (negative externality).
Let's see... reported total emissions are 40.5 GtCO2 in 2019 (38 in 2020). One GtC equals 3.67 GtCO2 so that's 11.04 GtC ... and the IMF says there's $5.9 trillion global "subsidies" for fossil fuels BUT the IMF gratuitously redefined the word "subsidies" to mean "subsidies and tax breaks plus negative externalities" and the actual "subsidies" part of that is only 8% of the total, but "social cost" obviously refers to negative externalities which is more like 90%... so my calculator thinks the negative externality is about $480/ton according to the IMF. So either they meant to say "$100/ton of CO2" rather than "of carbon", or I've made an error, or the US has a huge disagreement with the IMF...?
You said 2021 in a bunch of these, did you mean 2022? Or is it because statistics from 2021 are still not released and that's what you meant?
20% chance you will have a child, 11 months out, means either you are making a math error or you are trying to conceive. Assuming it's the second, so good luck!
EDIT: or you are thinking of adopting.
The contest is genius, thanks for running this!
Should "15. Bored Ape floor price here below current price of $203K" be $320K instead? That's what I see at https://www.coingecko.com/en/nft/bored-ape-yacht-club
I think ‘no reasonable person would invest in this’ is an incredibly low bar. Surely lots of intelligent people with money to spare make the occasional small moonshot.
Exciting that we can compete alongside your predictions this year!
A small suggestion: perhaps move away from using the Dow in future? It is a nonsense index with wacky methodology that made sense 100 years ago, and is mostly used by the media for historical reasons/financial illiteracy. S&P 500 (SPY) or total return (SP500TR) are much better.
Have you played D&D before?
I offer to be the dungeonmaster for an introductory session with you & 2-4 of your friends!
"supreme court will overturn roe v wade" - this may be difficult to adjudicate without clarification. By some standards it was already overturned (read: "clarified") by Casey. I suspect you mean, and it might be helpful to rephrase it to, "the supreme court asserts there is no constitutional right to an abortion".
You have two seemingly contradictory Taiwan answers. Clearly the betting markets need more participants so I could make money on arbitrage.
>95% chance they lose both houses of Congress implies 97.5% chance of losing each house, which seems too high
I don't think probability works like this when elections are correlated.
The actual math would be something like P(lose both) = P(lose House) * P(lose Senate | lose House)
Things are looking very dire for Bolsonaro around here right now. Most polls place him with around 25% of votes in the first round while Lula has around 45%. And the hypothetical rerun polls put Lula with a ~20% advantage too.
The economic recovery - which was Bolsonaro's big bet for a rebound on the polls - doesn't seem likely. Inflation is still above 10% and the increase of interest rates will most likely slow the recovery. Unemployment will probably remain very high (around 12%).
Bolsonaro's entourage was also betting that we would have overcome the Covid crisis and it wouldn't hurt the president's popularity as much anymore. However we are seeing that even with a much lower death and hospitalization rate, his approval won't go up. People are not forgetting his mismanagement (and straight brutality) during the pandemic that easily.
All in all, I would lower that 50% to about 20 or 25%. Bolsonaro will probably not get reelected.
Number 76 seems to be missing a percentage.
I hope this doesn't mean a 0% change of time off...
Covid "By current standards" question is quite vague. Are current standards the original two shots? Three with a booster? Four if you're Israeli?
It seems that poly people have a very different idea of marriage than I do.
Where can I learn more about what marriage is in the poly community?
10% chance Biden ends the year *not* as President? From both of you? What are your likely mechanisms?
Missing prediction: “Codex members have read at least one chapter of a new work of fiction I'm writing: [## %]”
I take it that second-guessing the Breyer retirement seems too easy now, but I’m thinking there is still room for a prediction. He has said it is contingent on getting a successor, and it’s not completely clear to me that this will happen.
is there an implicit 'as of jan 1 2023' for all of the non event ones? ( ie an approval rating, or a price point)?
How many subscribers do you have?
> 33. [redacted]: 80%
Sell to 40%, I don't think it's likely at all.
Hi, I'm the co-founder of Manifold Markets (with 30% chance of lasting through 2022 :P).
We've created prediction markets for all of Scott's predictions that were included in the contest!
Currently, you get 1000 Manifold play dollars if you sign up that you can trade with! It's fun, and the best traders will be tracked in the leaderboards for Scott's predictions.
If you want to participate in the contest, please do that first, and then come check out our markets: https://manifold.markets/fold/acx-predictions-for-2022
#52 and #53. Really?
How are we measuring prediction market size? I'm looking at alexa siterank, but that's only ordinal rather than cardinal so it doesn't make sense to talk of things being half as popular if that's the resolution source. Most of them don't publish anything about daily active users, total money transacted, etc, so the only metric I can find about all of them is alexa.
> 76. I make a time-off coverage agreement with someone
This one seems to be missing a probability?
I entered the contest. This is the full list of all our large disagreements (defined as one of us thinking either the YES or the NO side is >1.5x more likely than what the other thinks.)
Biden approval rating (as per 538) is greater than fifty percent (40)
25
PredictIt thinks Joe Biden is most likely 2024 Dem nominee (80)
90
PredictIt thinks Donald Trump is most likely 2024 GOP nominee (60)
75
Major flare-up (worse than past 10 years) in Israel/Palestine conflict (5)
8
Dow above 35K (90)
80
Inflation for the year below five percent (90)
75
Unemployment below five percent in December (50)
80
Google widely allows remote work, no questions asked (50)
35
Fewer than 10K daily average official COVID cases in US in December 2022 (20)
5
India's official case count is higher than US (5)
20
Some new variant not currently known is greater than 25% of cases (60)
75
Most people Scott sees in his local grocery store on December 31st are wearing masks (60)
90
Masks still required on US domestic flights (60)
90
China has fewer than 100,000 COVID cases this year (official estimate) (30)
70
New legal US real-money prediction market at least half as big as Kalshi (5)
10
New illegal but easy-to-use market satisfying the above (20)
30
Inflation for the year under three percent (FP: 80; SA: 80)
40
At least one country will have less than 10% of people vaccinated with two shots by November 1st (FP: 70; SA: 95)
99
A psychedelic drug will be decriminalized/legalized in at least one more US state (FP: 75; SA: 75)
50
The Biden administration will set the social cost of carbon at $100/ton or more (FP: 70; SA: 70)
35
Joe Biden is still president (MY: 90; SA: 90)
96
At least one Biden cabinet-rank official resigns (MY: 70; SA: 70)
85
>35. No current residents leave our housing cluster: 80%
Does this mean "nobody moves away" or "nobody moves away AND nobody dies"?
>9. No military conflict between the PRC and Taiwan (a worryingly low 90%) HOLD
My prediction last year (not sure if I posted it on here) was "30% nuclear war over the Far East within the 2020s". I'm not plugged into current events well enough to give a year-by-year. The implicit few% nuclear war this year seems fairly consistent, though.
Could we have the contest at the top next time so that we can give the predictions a shot before looking at your answers? As someone with no experience forecasting and knows that you have historically been well-calibrated, I feel like I should just copy your answers wholesale.
Viktor Orbán loses power - 60% is too much imo, FIDESZ has a stable lead in the polls.
“ 95% chance they lose both houses of Congress implies 97.5% chance of losing each house”
I don’t think that’s necessarily true, seeing as P(democrats Lose the House) and P(Democrats lose the Senate) are not independent. It seems like you could say P(Dems lose House) = 0.96 and P(Dems lose Senate I dems lose House) = 0.99 so P(dems lose Senate)*P(dems lose house) = 0.96*0.99 ≈ 0.95
Now that I’m writing these out, I agree that 0.95 is still to high, but its possible to have the conjunction of the two be 0.95 while they could individually be any number such that 0.95 < p(
X) < 1
4. Supreme Court will overturn Roe v. Wade (65%): SELL to 60%
There must be something I'm missing. This % seems really high given the base rate of overturning previous decisions is so low. I would have thought a straightforward overturning of abortion would ~30% likelihood with the more likely scenario(80%) they cripple roe vs wade.
Eternal glory is fine with me - but who wants "a free ACX subscription"? The point of subscribing ACX is the wish to show Scott how we really, really appreciate (I do) his blog by doling out some dough ( I did not. Yet.).
I tried to skew all my answers to the contest based on a few specific guesses about the future, rather than trying to be accurate. This should hurt my score in expectation, but won't it increase my overall chances of winning the contest by skewing my score distribution downward?
* I don't think I've ever been *explicitly* called a superforecaster, but SciCast once wanted to interview me because I was on their leaderboard. Does that count?
* What do you mean by "within two generations" exactly? Would a second cousin count?
Does "90% chance Biden is still president" means 10% chance Biden will retire due to bad health/death?
These jumped out at me:
"Biden approval rating (as per 538) is greater than fifty percent: 40%"
This strikes me as *very* unlikely. While better Covid situation might improve things, I wouldn't give this a better than 10% chance. *Obama* didn't have this after two years!
"Inflation for the year below five percent: 90%"
You should probably specify exactly which inflation measurement you're using here.
"China has fewer than 100,000 COVID cases this year (official estimate): 30%"
Huh. So *fewer* (even at just 30%) cases than in 2021, with Olympics and Omicron putting pressure on the situation? Well, I guess we're talking the official numbers here, so who knows?
Does "most likely" mean "more likely than any other option" or "51% or more"?
What are your predictions for mandates or rules only for vaxxed?
https://nakedemperor.substack.com/
Hmm, Main quibble points for me is on inflation (think chance of >5pct is higher) and unemployment (think >5pct chance is lower)
Stupid questions about predictions:
> 21. Starship reaches orbit: 90%
What exactly do you predict here? I could imagine at least three reasonable interpretations.
a) Starship is at an orbital height with velocity no less than orbital velocity at that height, and with direction of travel closer to horisontal than vertical, up vertically, and also very close to the pre-announced plan. This covers «clearly could go orbital but deliberately opted not to, presumably for failure impact mitigation».
b) Starship's position and movement are clearly orbital at some instant, even if an orbit is never completed.
c) Starship either completes a full orbit around the Earth above 100 km line, or intentionally releases an active payload that does.
> 29. Most people I see in the local grocery store 12/31/22 are wearing masks: 60%
Does this come with an implied «I go to the grocery store on 2022-12-31: >95%»? Or is it «last time in December 2022»?
> V10. China will not reopen its borders in the first half of 2022 (80%): BUY to 90%
What level of reopening? Business travel for non-state-level reasons in at least one direction, possibly with quarantine on top of quadra-shot and testing, or tourism in both directions with at most 7-day quarantine (with easily available level for US and China residents levels of vaccination and testing)?
> V17. AI will discover a new drug promising enough for clinical trials (85%): HOLD
For what values of discovery? Is something like https://www.science.org/content/blog-post/alphafold-excitement working out enough? (Human specified targets, specified search space, significant help from AI in filtering, estimated time savings in single-digit weeks)
> Y25. Russia does not invade Ukraine (60%) HOLD
I think it is a good idea to precommit to an information source choice procedure for the resolution. given that there might be pairwise difference between claims by Ukraine, Russia, and US. (It is also possible the difference won't be relevant for resolution).
Also, what level of sabotage/riot incitement infiltration as estimated by the chosen sources would count as invasion?
The #76 does not specify any percentage.
"9. Honduran ZEDEs legally crippled to the point where no reasonable person would invest in them further: 5%
10. New ZEDE approved in Honduras: 30%"
Given that I think no reasonable person should invest in Honduran ZEDEs anyway, this is going to be splenetic on my part.
So I think it depends how corrupt Honduras/Honduran government is, and the cursory Wikipedia search shows that it's pretty corrupt. After all, this is why ZEDEs picked Honduras in the first place; a poor country where the government would be happy to look the other way about setting up private fiefdoms so long as the right money changed hands "for the economic development of the area", naturally.
So if Honduras is *moderately* corrupt, I think that yes, there's a low chance on no. 9. If it's *very* corrupt (and the new guy seems every bit as scandal-ridden as his predecessors and is a member of the same party which has such lovely items on its record as "In May 2015, Radio Globo discovered documents that allegedly showed that the Honduran National Party had received large amounts of cash from nonexistent companies through fraudulent contracts awarded by the IHSS when it was run by Mario Zelaya. The contracts were approved by the National Congress of Honduras when Hernández was its president and the party funding committee was headed by his sister, Hilda Hernández. Hernández has accepted that his election campaign received money from companies tied to the scandal, but denies any personal knowledge. By June 2015, Hernández had appointed a commission to investigate the cause of the corruption." It's very convenient when you are able to appoint the commission that is investigating your own alleged corruption!), if it is *very* corrupt, then the chances go up.
Because ZEDEs are cash-cows for the governments of the countries involved, and what good is a cow you can't milk? The ZEDE for market gardening/horticulture should do okay, South American countries are familiar with the model of having peons to work on plantations for the benefit and enrichment of a handful of eminent and prominent families.
The ZEDEs which are "entice rich Westerners to come here and work remotely at their high-paying jobs while enjoying low cost of living and dirt-cheap local labour which will have no rights under local law since *you* will be the local law, hence every stakeholder can be lord of the manor" are going to be milked *hard* because you're not producing anything locally like the giant glasshouses/polytunnels ZEDE, and that will mean all kinds of levies, taxes, charges and other "Hello, I am the Minister for Brown Envelopes, so long as you keep shovelling money into my private Swiss bank account your holiday resort will be left alone". No, I can't put a number on this. Yes, I think this bubble will eventually burst.
"4. Supreme Court will overturn Roe v. Wade (65%): SELL to 60%"
They're giving as high as 65% for this? Well, given that this is a perpetual scarecrow for liberals ('they are going to go after abortion rights!!!!' every time there's a Supreme Court nomination coming up), I'm not that surprised that Vox are getting ants in their pants about this, with the Breyer announcement and a majority Catholic court (and believe me I am immensely tickled about the spectre of the Horrid Popish Plot being dragged out of its grave once more).
I don't think it's going to happen. Things have moved more on rolling back abortion than I ever thought would happen, see the Texas laws, but there was swift movement to combat this, see the Texas laws. So I'd put it more around 40%, if even that high. A possibility because it's always a possibility, but nowhere near a likelihood.
"5. Stephen Breyer will retire from the Supreme Court (55%)"
I thought he said he was definitely going? Especially since his own party has been handing him his hat and showing him the door? https://blogs.lse.ac.uk/usappblog/2022/02/01/with-breyers-retirement-the-us-congress-has-lost-its-champion-at-the-supreme-court/
"21. Kenneth Branagh's Belfast will win Best Picture (55%)"
Ooooh. This one. I was only very vaguely aware he had a new movie out, and looking at the synopsis. Well. Yeah. Standard Troubles movie. I don't know if it's strong enough to win, or what movies it is going up against. I would say 'no' for various reasons (e.g. it's too white, not diverse enough, for modern tastes) but maybe it will be seen as a safe bet to vote for because it's not American?
The real problem is that this year is the 50th anniversary of Bloody Sunday, something which is still very raw and still affecting relationships between Ireland and Britain, and "Belfast" seems a bit too *safe* a treatment of the whole tangled mess:
https://www.irishtimes.com/news/crime-and-law/bloody-sunday-what-i-remember-most-is-the-whistling-of-bullets-1.4789410
For a start, this is Branagh reflecting on, or presenting a version of, his childhood in Belfast. Bloody Sunday was in Derry. It's set at the start of the real breaking out of violence, before the British Army becomes the villain of the piece. It's on a personal level of a family and a small boy. It's extremists on both sides who are the problems, with a sort of "why can't we all just get along?" message. It may appeal to voters as a safe choice that deals with Issues that are in the past (the 60s) and far away (Ireland) and so they can have their cake and eat it.
I don't think it will win, though. I think it'll get some awards but not Best Picture.
While we're talking about the North:
"27. Sinn Fein becomes the largest party in the Northern Ireland assembly (60%)"
At the moment, the two major parties are the Democratic Unionist Party with 27 seats and Sinn Féin with 27 seats, the remainder made up of odds and ends:
http://aims.niassembly.gov.uk/mlas/statistics.aspx
It looks like they'll pick up some extra votes, but the thing is that the parties cannabalise each other; for Sinn Féin to get extra votes, they will peel those away from SDLP and smaller parties. The DUP will do the same by absorbing Unionist votes and replacing, say, the 1 seat held by a small party candidate. As you can see from this news story, where the 1% for Sinn Féin has come from the SDLP losing 1%:
https://www.rte.ie/news/ulster/2022/0122/1275237-sinn-fein/
"The gap has widened since the last LucidTalk poll in November, with Sinn Féin up one point and the DUP down one.
Of the other main Stormont parties, the poll puts Alliance and the Ulster Unionists in joint third place on 14% (Alliance down one point on November and no change for the UUP), the TUV on 12% (up one) and the SDLP on 11% (down one)."
Can you compete if you already looked at prediction markets before you posted, and have seen many of the questions, but don’t look while making your submission?
Yglesias #2. Net or Gross?
If you'd like to play D&D, I'd love to run the game for you and a friend. I've been DMing a while, and I've been interested in your writing since you were posting on LW as yvain. A lot of the themes in your work (particularly The Goddess of Everything Else) touched me deeply and affected my own writing (I wrote Significant Digits, the HPMOR sequel). Happy to do a one-shot trial first to see if the vibe works.
60% chance of Donald Trump being the GOP candidate? So that means you think there's a 40% chance he will die in the next 2 years? Because that's the only way I see him failing to get the nomination, such is his grip on the party.
(28) seems like it might need a little more clarification? Does a "new variant not currently known" need to be one assigned a new Greek letter, or have some number of mutations relative to current variants? The media's been very excited lately over possible vaguely-variants like "delta-cron" or "omicron 2" or whatever, so I can easily see a future where something the media calls "mega-delta-omi-doomy-zord" or something is responsible for a large number of cases, even though it isn't technically a distinct variant by the standards used by official health agencies.
The odds for democrats holding one of two houses seems way too low. Currently betting markets only have republicans with 85% chance to win the house and 75% chance to win Senate. So 90% chance they win both seems overconfident.
The temperature forecast also looks overconfident towards warmth. In the last 30 years only 60% of years have been warmer than the year before them. Even in the small sample of the last 10 years its only happened 70% of the time.
US/WORLD
1. SELL to 20%. Biden is currently at 42% (as per 538). Getting to 50% would mean improve his polls by +8 pp. This is better than any president has done between the end of their 1st and 2nd year since Truman. Most presidents' approval ratings decline during this time period.
9. BUY to 15%. The recently elected president explicitly campaigned against them. The ZEDE law is in the Constitution, which should make it hard to overturn, but not nearly as hard as it would be to change the US Constitution. How would this resolve if the government passes (and enforces) a law which blatantly contradicts the Constitution?
ECON/TECH
19. BUY to 60%. Unemployment is currently 4%. So you're thinking that the unemployment rate is more likely to go up. 4% is unusually low, but the unemployment rate usually trends downward, and only occasionally rises rapidly. I also don't see how this is consistent with Biden's approval rating increasing +8 pp 40% of the time.
COVID
You have predictions on how high the floor will be for endemic covid, but not for how high the ceiling is. This seems like an easier and less useful prediction.
25. BUY to 10%. So far in the pandemic, India's case numbers have been higher than the US for slightly more than 2 months (May & June 2021). This is about 10% of the pandemic. The US also has a smaller population, is better vaccinated, and probably has had a higher percent of infections.
28. BUY to 80%. This has happened at least 2 times per year so far in the pandemic. I also think that this is more likely to happen than WHO designating it as a variant of concern (VOX 13) - especially since the new variant would have to be more infectious, but not necessarily more dangerous.
29. I'm not familiar with where you live, but I hope this isn't true. I suspect that this will be true for most of the country by the end of March. Which is a long time before December.
32. SELL to 20%. This will only get harder with time. New variants will be more infectious. There will be more pressure to open borders as the rest of the world returns to closer to normal.
BLOG
77-79. It's interesting to get an order of magnitude estimate of how much this blog makes.
82. BUY to 70%. Even higher if different cities in the Bay area count as more than 1.
87. This is a "Science" blog? There isn't even a Molecular Monday for new drug tests.
VOX
3. Don't you have a 50% chance that unemployment will go up to 5%?
10. This is necessary if they're going to keep a zero covid strategy.
15. There are currently 36 countries with less than a 10% vaccination rate (as per NPR). North Korea is "No Data" and so isn't included. https://www.npr.org/sections/goatsandsoda/2022/01/14/1072188527/for-the-36-countries-with-the-lowest-vaccination-rates-supply-isnt-the-only-issu
16. Is this for an entire state? Or does a city in a new state count?
18. That's depressing.
19. SELL to 40%. The current social cost of carbon is at $51/ton. Is there a reason why Biden would double it? One of the first results I found from Google is a Vox article arguing for $258/ton ... by one of the authors of the predictions. https://www.vox.com/future-perfect/22643358/social-cost-of-carbon-mortality-biden-discounting
22. Sell to 40%. Norway is probably the favorite, but I don't think that any country should be above 50%.
YGLESIAS
22-23. Inflation is often a self-fulfilling prophesy. If people think it will happen, it will, and if people think it won't, it won't. Vox is much higher profile than Yglesias. Vox doesn't want there to be high inflation, and so might purposefully underestimate it, in the hopes that they can influence it. Yglesias is low enough profile that he probably doesn't think that his prediction will impact what happens. I would lean towards Yglesias's prediction.
Assuming I wanted to duplicate Scott's bets because I revere him as an insightful man of foresight and wisdom, where can I do that? I'm moderately familiar with the concept of prediction markets, but I've never placed a bet.
Much too bullish on inflation. Doubt very seriously it will be below 4-5% for the year.
Also, your triple prediction on COVID treatments seems way too low at 1%. Would go with 25%.
I'm confused by the inflation ones.
(on own predictions)
> 18. Inflation for the year below five percent: 90%
(under Vox)
> 2. Inflation in the US will average under three percent (80%): HOLD
(under Yglesias)
> 22. November 2022 year-on-year CPI growth is below 6% (70%) BUY to 80%
> 23. November 2022 year-on-year CPI growth is above 4% (70%) SELL to 50%
Isn't CPI growth the same thing as inflation? Or at least, it's one of the standard measures of inflation. It seems odd to have inflation under 5% but CPI above 6%, and yet you're assigning at least a 10% chance of that. Similarly, it seems odd to have inflation under 3% but CPI growth above 4%, and you're assigning at least a 30% chance of that.
I have no real comment on the predictions, but I have to laugh darkly at the phrasing "at least $250 million in damage from a single round of mass protests".
Traditionally, one would not refer to events causing upward of $250 million in damage as "protests".
"Viktor Orbán loses power in Hungary (60%)"
I don't know what "loses power" means here precisely, but a 60% estimate of him losing the election is wildly off. In my experience, basically every sensible person in Hungary thinks that a small-margin Orbán win is the most likely scenario. I mean, it's not impossible that he loses, but it's surely below 50%.
As a brazilian , I can´t agree with this bullish analysis of Bolsonaro´s chances of reelection , 50 % seems a lot right now. He is pooling worse than any other who was able to be re-elected (i.e: Temer does not count). I would to read counterarguments
“…based on Vox’s 22 Predictions For 2022 and and Matt Yglesias’ predictions…“ has two “and”s in it.
"at least $250 million in damage from a single round of mass protests"
How is "single round" defined?
Can that be n days in m cities with n, m > 1, assuming they all those individual protests share the same cause (assuming there is a specific one, that can be pointed to)?
Yglesias is too optimistic about inflation. My predictions for "Consumer Price Index for All Urban Consumers" change over 2022:
≥5% increase 90%
≥6% increase 70%
≥7% increase 50%
≥10% increase 20%
This inflation is caused by Trump and Biden printing money to pay for the Covid stimulus at the same time that overall production fell, and it's going to take at least a couple of years to work through the system.
Surprisingly, in my 5 minutes of research I adjusted way down on Yglesias 2 and a little up on Yglesias 3. I was expecting to agree with 80% since it's likely to be a red year and the Senate skews red anyway, but the particular seats up for election look good for Democrats. They aren't defending any Trump-won seats, and I expect these midterms to lean red rather than be a red wave. That means the states I judge to be at risk of flipping are Georgia, Arizona, Colorado, New Hampshire, and Nevada. I didn't have time to look up each individual approval rating, but if I remembered right, most of these Senators are reasonably popular in their states. I think Georgia is more likely than not to flip, but the rest are (individually) more likely than not to remain blue. Altogether I do think at least one will flip, but I wouldn't put my confidence at 80%.
As for Yglesias 3, the next-reddest states after those 5 are Illinois and Washington. Even in a wave election I don't see those flipping. 80% is too low.
76 is missing a number
77-79 Per year? Per month? Per second? Projected over its entire future existence?
82 Is that 6 meetups per city, or 6 meetups total? If total, would it change the meaning of the sentence if you just deleted the first "6"?
As long as people are hanging around here forecasting, might as well enter the 7th annual version of my contest, which looks much like this: https://forms.gle/P3tR8xccsYjYTi5A9 (late entries are fine, just not eligible for the prizes: a book and a bowl of pho)
how is >66% US population fully vaccinated (by current standards) against COVID being evaluated? does this mean Jan 2022 standards (not boosted), Jan 2022 standards (boosted), or the standards in Dec 2022?
I like how Yglasias specified CPI when talking about 'inflation'. i know there is a general understanding that when officaldom talks about inflation they mean the CPI, but they are truly not the same thing.
I find many econ ideas are purposefully or needlessly convoluted and looking into them to unpack what they are saying means a lot. Out of the three big ones I hear most often, GDP, Unemployment, and Inflation....the offical numbers are representing a very narrow and very politically driven set of figures which get reported.
I'd encourage anyone to research further into what is and is not included in the CPI number and see if you think that value represents the inflation you experience in your own life.
It does not include any information about the money supply or real estate prices or industry. It also uses a crazy substitution principle to try to 'mimic' what real world buyers might do.
So if the price of beef goes up and chicken is cheaper, they'll assume for you that you'll switch to buying chicken instead. So.....this number is NOT telling you about the changes in the price of goods then. It takes real inflation and a real upward movement in prices and finds a way through biased arbitrary selection to remove that inflation from their calculation of inflation.
Same with electronics which are one of the worst offenders. Buying a new phone or computer after 2-3 years? Say you're a 'I want the flagship phone' type of buyer...If that phone was $1,500 a few years ago and is $2,000 today, then the price you'll pay has gone up for 'the same' product of the flagship phone.
But they'll pretend that the new 2nd or 3rd tier phones are equally if not more powerful than the ones from a few years ago or have longer battery life or any number of arbitrary selections they make to pretend you can now get 'the same' phone for $800 instead. Every single year for around 20 years they've had this nice negative figure to modulate the overall basket in any way they wish using whatever choices they want to make that year in officialdom.
If you have this much freedom of choice or degrees of freedom in what you're doing...then you can say almost anything you want to say. In a deep dive on this, even 10% of this level of freedom of choice would see Scott lambasting a psychiatric drug study and saying how we'd be utter fools for using it or relying on it to tell us about that drug. They can pick and choose every single variable and arbitrarily include or exclude things such that every year the CPI 'metric' is measuring a different thing while attempting to lie to you about how they're measuring the same 'conceptual' idea of their guesstimate of consumer choices...a concept whose components are redefined every time.
And do we go back and update the CPI to a 'real one' after we interview lots and lots of people to get a refined measurement of consumer behavioural changes in buying choices? Nope! There is not even a hint of a notion that officaldom would ever do that with their CPI.
They are not doing science or the lower religious artform of economics, they are not seekers of truth or trying to tell us something worth knowing about the world. The lack of updates, accuracy, or even an attempt to do this tells you all you need to know. The official CPI is a political and propaganda tool to influence financial markets first and foremost.
It is a pretend game of counting and not counting major categories of prices which affect people's lives...real estate up 30% in your area while Blackrock goes on a buying spree spending 20-50% above market rates? Buying a home is the single largest purchase in the average person's lifetime....don't worry! We DIDN'T count that in 'inflation' or CPI specifically.
Truly this is inconceivable! In the classic meme sense of Princess Bride 'I do not think that word means what you think it means'
CPI is so wonky, manipulated, politically driven, and diverts so very very far from reality and what people's lives are like that I'd assert it cannot be anything other than a lie and is a part of the propaganda machine more than it is is part of any academic or trustworthy way to attempt to describe changes in the real economic landscape.
which grocery store do you shop at? I think the P(most customers masking in Safeway) << p(most customers masking in BiRite/Berkeley Bowl)
against FP you have Emmanuel Macron at 65% but against MY you only have 60% (technically I could see it the other way because MY only says reelected but FP has reelected president of France so it should be X% and X.0000001% where EM is appointed to some other position and then runs and wins reelection)
I would like to see some more skin in the game predicting about education around here. Too much Bryan Caplan and Freddie DeBoer (sorry guys), plus motivated reasoning inform the discussion. As a first step in a remedy to this here is a market that is OBVIOUSLY mispriced. https://manifold.markets/JohnBuridan/will-any-us-state-excel-its-previou
The Metaculus team is always excited to see Scott's yearly predictions. We actually made a short forecasting series featuring a number of the above questions if folks would like to predict along with them: https://www.metaculus.com/questions/9680/scott-alexanders-2022-predictions/
Fun! for the betting market, I'm just following Scott.
Except for the chance of kid, If he's really trying to make a baby, I'm saying ~80%, but I don't know that's true so I'm going to go 50% of making a baby... which is betting with my heart. :^)
Having a kid changes your whole outlook on life. It's magical and a pita.
maybe off topic, but not sure a better place to answer a q i've been exploring a while: what is the best corporate prediction market software out there? looking for something cloud-based to build out with teams, and my hacked version in google sheets doesn't exactly inspire confidence or confidentiality.
Kudos for putting predictions out there.
But leave out the personal ones. No way to independently judge them and they are irrelevant to anyone else but you anyway.
I see them mostly as padding your success rate under the presumption that you know you.
You have inconsistent predictions for Macron being reelected, with one at 65% and the other at 60%.
Very Interesting. I also made the same predictions for the contest.
There are a few areas where I disagree to a major extent with the predictions offered by Scott Alexander, Vox Future Perfect and Matt Yglesias. Here is a quick rundown (those predictions with more than 25% difference)…My predictions are in the row below the original prediction.
- PredictIt thinks Joe Biden is most likely 2024 Dem nominee (80)
My prediction: 40% - I feel like if Democrats loose the midterms, then Biden will just resign on his own and Harris will step up. Also, other Democrats could be making a bit of a fuss about him if he doesn't deliver...
- Dow above 35K (90)
50%: Already, the Dow is showing signs of struggle. If the Fed raises rates, then it's hard to see the Dow to continue to being as strong...
- Inflation for the year below five percent (90)
50%: That was already said last year that inflation would only be "transitory"...well that wasn't the case ....so why should it change this year...only if the Fed raises the rate quite a bit...but will they?
- Starship reaches orbit (90)
50%: To be honest, I don't know much about it, but considering Musk's track record in getting products to the markets in a timely fashion (Roadster?), I don't think it's that realistic...
->66% US population fully vaccinated (by current standards) against COVID (70)
95%: Didn't that already happen? Anyway, looks to be veeery achievable …
- Masks still required on US domestic flights (60)
30%: I don't really see the point in this...if there are tests before every flight, and/or vaccine mandates, then why insist on masks (especially if they are just surgical masks?).
- New legal US real-money prediction market at least half as big as Kalshi (5)
50%
and -New illegal but easy-to-use market satisfying the above (20)
50%: I don't know much about these, so I just said 50%...
- Inflation for the year under three percent (FP: 80; SA: 80)
10%: That seems like a ridiculously high number...I think it's really unrealistic to have substantially lower inflation this year...oil prices are still rising, as are many other input prices for many goods...
-Democrats will lose their majorities in the House and Senate (FP: 95; SA: 90; Matt Yglesias also predicted this event, saying 90)
50%: Yeah, like I said this seems to bullish on the Republicans...they probably will win the house, but both the house and senate? That didn't happen in 2010, and while Biden is now less popular than Obama was back then, the difference isn't that huge, and Republicans are now much more of a non-mainstream party...
-Supreme Court will overturn Roe v. Wade (FP: 65; SA: 60)
40%: This also seems rather high...I mean it's been the law of the land for so long...don't think most judges would want to change it, and opening up a whole new can of worms...
- Jair Bolsonaro will be reelected president of Brazil (FP: 55; SA: 50)
30%: Well, he's badly trailing in the polls, and seems to be very unpopular...of course they said that about Trump, but Lula is no Hillary...so except if he rigs the election, or Lula gets jailed, I don't see him winning...
- China will NOT reopen its borders at any point during the first half of 2022 (FP: 80; SA: 90)
60%: I am not sure...how long can they really go with this strategy? Probably this year, but then they would have to rethink if it's not a net disbenefit...
- 20% of US kids between 0.5 and 5 years old will get at least one COVID vaccine by year's end (FP: 65; SA: 65)
40%: Don't think it's that realistic to be honest, as so many anti-vaxxers are people with children...also, maybe there won't be as much pressure for people to get vaccinated anymore?
- A psychedelic drug will be decriminalized/legalized in at least one more US state (FP: 75; SA: 75)
50%: Hard to say, but I would be less bullish on it...Cannabis yes, but these types of drugs (LSD?) seem not to be as popular with the broader public so IDK...
- AI will discover a new drug promising enough for clinical trials (FP: 85; SA: 85)
50%: I don't think AI is advance enough....
- Norway will win the most medals at the 2022 Winter Olympics (FP: 60; SA: 60)
30%: Not sure why a small country like Norway should be the favorite? I mean Russia, the US or even Canada sound more like favorites to me...
- Democrats lose at least two Senate seats (MY: 80; SA: 80)
30%: Which ones? I think they might loose NH, maybe GA but probably not AZ... Though I think Democrats can win PA, and possibly WI and/or NC...so on net its -1 to +1, so 0 net seats on average...
- Liz Cheney loses primary (MY: 80; SA: 80)
50%: I don't know...possibly, but I would give her higher chances than this though...
The predictions from Macron seem ridiculously low. I've seen an analysis from the Economist (https://www.economist.com/interactives/france-2022) that gave Macron 90% odds of winning, and that matches both my intuition and the polls. His only remotely serious opponent is Pecresse, and he's still projected to win against her (I'm really puzzled anyone could prefer Pecresse over Macron, but I could see Pecresse winning). Everyone else has no chance.
I'm not into prediction markets, but I'm tempted to actually get in on this one. Either the market is seeing something I'm not or it's suffering from american-centrism and guessing wildly off the mark.
Inflation <5% at 90%. Really? I would not dare to put that higher than 50%. It is already spiking at 7% right now. Labor shortages, oil market is very tight with limited drilling inventories in US shale fields at lower break even prices. And supply chains will see more pain with China sticking to zero Covid in 2022.
If you put a major flare up in Ukraine at 50%, then that alone could push inflation above 5%. Since Russia is a major commodity supplier. And any interuptions to export could send commodity prices flying.
Nobody seems to have mentioned this: at the beginning of 2018 (slatestarcodex.com/2018/02/15/five-more-years/) you gave Roe v. Wade being overturned a 1% chance. Now you give it 60%. Regardless of what actually happens, at this point you'd have reduced your money to a 60th of the total if you'd bet on this.
I'm amazed anyone would ever give Roe being overturned a 1% chance, let alone under a Republican president and senate, with 2 or 3 liberal or moderate justices soon likely to die or retire, let alone someone as good at prediction as you. I guess I'm just asking (light-heartedly, with no offence intended): what on earth were you drinking when you made that earlier prediction?
You make $400000/year with the blog? Awesome, I am impressed! But how is this possible?
> Dow above 35K: 90%
A bullish position given that the Dow has been hovering around 35K recently. Why so much confidence?
> No recession in 2021 (90%) SELL to 80%
Hmm, what kind of bull is this?
> Medical establishment reverses course and officially says any of Vitamin D, HCQ, or ivermectin is actually effective against COVID: 1%
Also surprising. Earlier you gave a 10% chance that ivermectin works ... is there new data that made you update even more against ivermectin? Or is it just that you think they won't "officially" support it unless the effect size is shown to be unexpectedly large?
> Fewer than 10K daily average official COVID cases in US in December 2022: 20%
...for reference this would be below the 2021 low point on June 20, 2021, and below the low point during the 2020 lockdowns. So, it's obviously unlikely: it would require no new variant-of-concern to arise, and it would also require immunity that endures long enough to prevent flare-ups. Alternately, a genuine apocolypse could cause this result, or a halt to the gathering of statistics.
> At least three dates with a new person: 30%
This I did not expect, Mr. Groom, especially since you said in 2014 that you were asexual.
> The Biden administration will set the social cost of carbon at $100/ton or more (70%): HOLD
Wait, what effect does this number have? I thought it seemed very high as a carbon tax, but low as a estimated social cost (negative externality).
Let's see... reported total emissions are 40.5 GtCO2 in 2019 (38 in 2020). One GtC equals 3.67 GtCO2 so that's 11.04 GtC ... and the IMF says there's $5.9 trillion global "subsidies" for fossil fuels BUT the IMF gratuitously redefined the word "subsidies" to mean "subsidies and tax breaks plus negative externalities" and the actual "subsidies" part of that is only 8% of the total, but "social cost" obviously refers to negative externalities which is more like 90%... so my calculator thinks the negative externality is about $480/ton according to the IMF. So either they meant to say "$100/ton of CO2" rather than "of carbon", or I've made an error, or the US has a huge disagreement with the IMF...?
> 2022 will be warmer than 2021 (80%): HOLD
Hmm, not unreasonable. Even though there were some huge heatwaves in 2021, especially the record-smashing Pacific Northwest heat dome, 2021 was still a La Niña year (https://psl.noaa.gov/enso/mei/) and below trend (https://crudata.uea.ac.uk/cru/data/temperature/). But note that the latest expert prediction is that there is potentially another La Niña this year: "La Niña is likely to continue into the Northern Hemisphere spring (67% chance during March-May 2022) and then transition to ENSO-neutral (51% chance during April-June 2022)": https://www.cpc.ncep.noaa.gov/products/analysis_monitoring/lanina/enso_evolution-status-fcsts-web.pdf
Thus, global warming is doing a nontrivial amount of work in this prediction. I endorse it though.
> Bongbong Marcos will be elected president of the Philippines (55%): BUY to 60%
Am wondering how you came to have any opinion on this.
11. Saudi Arabia and Israel establish diplomatic relations (60%) SELL to 50%
Israel's FM and PM-in-waiting is Yair Lapid. He anagrams to "Paid Riyal". This is not a coincidence, because nothing is ever a coincidence.
Looking back on 2022, most of this was not on the mark..