Claiming to be vegetarian and being vegetarian are two very different things :P

Some surveys find that >60% of vegetarians have eaten meat the very day before they were asked whether they are vegetarian.

https://www.econlib.org/archives/2013/07/vege…

Not a big fan of claiming = being, but I suppose you can just claim the question doesn't care about actual adherence and just cares about claiming.

@Jgalt What Trump says and what he does (like it or not) are very different things :P

I'd really like more short-term predictions. I don't want to wait 10 years for my predictions to resolve :P

These betting sites seen overconfident, especially given some local reports coming in from 538 about how absentee ballots are changing the totals. 30-40% chance seems reasonable to me at the moment. Remember that mail-in votes will favor Biden.

Seems like he's running at 0.97/week now

Probability of this event has gone up in the community as his 538 chances have gone done. Weird.

Bezos is worth about 150 billion right now. Assuming he has all his money in the market at 8% return per year, that means in about 25 years he'll be worth a trillion: 150000000000*1.08^25 Uncertainty: if most of his money is in Amazon stock (which it seems to be more true), we need to use Amazon's growth rate instead, which is kinda astronomical: about 7x over 5 years. This would reduce the time needed. But it's very hard to predict that Amazon will grow at the same rate forever: E.g. MSFT peaked in 2000 and it took around 14 years for it to return to t...
@(shiranpasternak) wrote: > One reason for posing this question early on in the first administration is to track how the prediction market reacts to news. It seems to me that the [NYT exposé on Trump's wealth](https://www.nytimes.com/interactive/2018/10/02/us/politics/donald-trump-tax-schemes-fred-trump.html) published yesterday — which reads more like an indictment of a tax system favoring the top 10% than proof of nefarious behavior by a specific wealthy family — will only serve to rile up and entrench the Trump base. I've adjusted my prediction accord...

I'm guessing we will be able to resolve this question sooner rather than later. I don't think anyone will claim the US is in a recession as of Jan 1, 2020. I still think there are some troubling signs over the next 6-12 months (such as a slight dip in private investment), but so far, no recession!

Another data source:

https://www.kansascityfed.org/publications/re…

Seems like since '85 there have been only 2 earthquakes in the US with > $10 billion in damage. This implies a 6% yearly chance of such an earthquake. Over 5 years, the chance of >=1 such earthquake is therefore about 26%.

Why was the Metaculus prediction so high and trending up at the end?

It seems that, historically, ~44% of relevant presidents are reelected. By relevant I mean presidents who neither died in office nor refused to seek reelection. I also exclude the first 4 presidents who faced mainly token opposition: https://www.thedailybeast.com/for-us-presidents-odds-for-a-second-term-are-surprisingly-long (I don't know why the article includes presidents who died or didn't seek reelection in their 30% figure). Is this a good base rate? Ehhhh, maybe as good as any. Given that Trump's favorability is close to that number, it seems re...

@Anthony wrote:

Because there are two questions on this, but both have closed, I've re-opened this one for prediction, so that it covers the time between the Nov. election and (nearly) the cutoff date for Senate confirmation before the next congress.

I'd have opened a new one. It's worth having the initial prediction set in stone and opening a new one, in my opinion. That is, I think it's good to punish predictors for missing low-probability events by question closing time.

So hard to predict, such wild swings... Here is a chart of the relative market caps: https://i.imgur.com/h9XNxdg.png We hit >80% once in 2017. If we say there's a 50% chance of hitting the same number next year and at that point a 80% chance of surpassing it by enough to win, that's a 40% chance of winning in the next year. We have almost 2 years left. Just based on this data, I'd say chances sit at roughly 60%. This is a very rough and stupid estimate, of course, but it gives a starting point from which to adjust. Google trends shows about 6-10x more...
@(Anthony) wrote: > @Fruo By default, if the resolution criterion has been satisfied prior to opening, the question resolves ambiguous, per the [faq](https://www.metaculus.com/help/faq/#whatifres2). However I think in some cases, where the question is about a potentially-repeatable event with some verb tense ambiguity, we could all agree to keep it alive and look for the next instance that occurs after question opening. Hmmm, interesting. I never knew I should read the FAQs. They should be renamed to "Mandatory rules of the game". I say this tongue in ...

@Matthew_Barnett But is that how it's worded? "Until X Event" seems to imply until the start of the election. So maybe this should have been resolved when the first primaries started?

Starting with the base probabilities, since 1947 there have been 11 recessions. If we assume the average recession lasts 1 year (seems about right), that's 11/71 recession years = base rate of 15.5% of years are recession years. I think this gives a roughly 29% chance of any two random years having a recession. We can now start adjusting from that base rate. 1 month of those 2 years has already gone by and we're likely not in a recession. Looking at graphs of unemployment, it seems that it ticks upwards slightly right before a recession starts, so if we...

I'd like to note that about 10 days after resolution time, the ratings are no closer to the target:

https://projects.fivethirtyeight.com/trump-ap…

— edited by Fruo

A useful exercise is to try to replicate my analysis for a similar period of time before the 2008 recession (22 months, roughly, so February 2006): At the time, we had had 10 recessions over 59 years, so that's a 16.9% chance of being in a recession on any given year. This gives a 31% chance of recession by Dec 2008. The BCIg at the time was 74% of what it was now. This reduces my correction for its health indication from 38% to 28%. So far, no great change in analysis. However, when we get to the 10-2 treasury spread, it sat at about -0.1. From past ...

Why are the forecasts here so different from oil futures? Oil futures predict a roughly $56 price in December 2018: http://www.cmegroup.com/trading/energy/crude-oil/light-sweet-crude.html

Even if you want to add in the prevailing interest rate to it, we get nowhere near the median prediction of $63 on here. And the 50% spread on our prediction is $35?? Seriously?