Will I believe the AI risk letter markets were priced rationally?
23
660
470
resolved May 31
Resolved as
80%

Resolves to a percentage score of how rationally I think the linked markets were priced, based on my extremely subjective opinion. Resolves after sufficient relevant information becomes available to me, or after I believe it's unlikely that more relevant info will become available.

Clearly the price is very volatile, so I'm not going to treat that as irrational for the purpose of this question - I'll look at basically a moving average of the market price.

I will not trade on this market.

The markets:

https://manifold.markets/quinesweeper/will-there-be-another-wellrecognize

https://manifold.markets/quinesweeper/will-there-be-another-wellrecognize-511a499bd82e

https://manifold.markets/quinesweeper/will-there-be-another-wellrecognize-393de260ec26

Get Ṁ200 play money

🏅 Top traders

#NameTotal profit
1Ṁ404
2Ṁ84
3Ṁ46
4Ṁ27
5Ṁ22
Sort by:

Based on the information I have so far, which includes the public commentary and some private conversations with people involved in the market, I believe the markets for by June/July were priced reasonably, although slightly underpriced, while the by May market was substantially underpriced for most of its duration. Even considering that the well-informed traders were understandably keeping their information hidden, in my view the market equilibrium "should" have been more like 75% than 50% and the market could very plausibly have traded there if the market had functioned more effectively. The May market probably should have looked more like the July market, which mostly stayed around 75% with a lot of volatility, and the July market could have gone even higher except people (rationally) traded much more on the earlier markets first.

I think there was enough information to strongly weight the evidence of insider information much more than the base rates. I personally had predicted 90% by July and 55% by May - I was mainly uncertain about the timing, not about the existence of the upcoming letter. In retrospect my May prediction probably should have been higher as well, I made some avoidable mistakes in my analysis of the situation.

The reason for the underpricing was a combination of several factors. I think the most important being:

  • misleading commentary on the market - certain traders made comments about the reasoning behind their trades that were not lies but were highly misleading, whether intentionally or unintentionally, and those substantially influenced the predictions on the market.

  • the memes were a contributing factor to the uncertainty and volatility on the market, and made it harder to accurately infer what was going on - which was definitely in part intentional by some of the YES predictors. They didn't affect my prediction much, but I suspect they influenced some people to bet a lot more NO than they otherwise would have.

  • keeping information hidden - even leaving aside the firsthand insider information which was not revealed because it was under embargo, there was plenty of secondhand information that could be deduced, and some of it was deduced by some of the traders. But markets disincentive sharing this sort of information, and in this market more than most because when there's hard news or interesting analysis people often share it, but in this case the relevant information that potentially could have been shared mostly took the form of more speculative analysis of other people's trading, not of the ground facts.

I'm taking as a given that markets aggregate hidden information, I'm not saying the market should have gone to 100% - just that the equilibrium price could and probably should have been higher, and that uninformed traders betting on base rates likely should have updated more on the behavior of the market (even taking into account that the market was full of many meme traders).

I'll make a decision about how to score this market, I'm roughly thinking 8/10 based on the May market being substantially underpriced and the June/July markets being fairly reasonably priced.

bought Ṁ50 of NO

Does this have any real chance of resolving YES if the original market resolves NO?

@NcyRocks Sure it does, if I believe in retrospect that the market price was a reasonable prediction.

bought Ṁ0 of YES

@NcyRocks if a letter comes out in June, it wouldn't seem so unreasonable

@finnhambly Yep. A letter coming out in June would be strong evidence in favor of YES

predicted NO

@finnhambly Good point