[tptacek]

The idea is that people have all sorts of fragmentary information about future events that they can't directly reveal, due to confidentiality or trade obligations (among other things), and that a prediction market effectively liberates the directional content of that information by converting it into prices.

Robin Hanson can credibly claim to have invented prediction markets as we understand them today.

[giarc]

"that a prediction market effectively liberates the directional content of that information by converting it into prices."

I can see this, and I guess maybe my issue is with the phrasing of "aggregating" insider information. Because you aren't just aggregating insider info, you are also aggregating non-insider information, but no one (but the insider) knows what is right.

Is there different types of prediction markets then? One where there is a true insider and one without? For example, you could take bets on weather it will rain on Saturday. People can make educated guesses, but no one really knows (no insider). On the flip side, Kanye could create a bet on whether he will run for president. He would be the only insider, so again, aggregating insider and non insider information.

[kelnos]

> you are also aggregating non-insider information

You're not really aggregating non-insider information, because in these cases, it's not really "information", it's just (at best) rational guessing or (at worst) gambling.

But yes, Kalshi and Polymarket essentially aggregate gambling, rational guesses, and insider information that's likely to be correct. It's a losing game unless you're an insider, and these companies profit off of other people's addictions.

[tptacek]

I would argue that a "market" in whether it will rain on Saturday isn't really a prediction market, or even a market, at all. It's just a bookmaking operation. The core function of any market, in anything, is price discovery.

What's the difference between a "Monday it will rain" market and a NCAAF prop bet on a team's rushing yards? I could argue that DraftKings prop bets are actually more like prediction markets than these "will it rain" bets. People actually do have directional information to contribute to sports propositions!

(I think online sports betting is evil.)

[kasey_junk]

Tradeable risk is the difference between the ncaaf bet and rain futures. Levine has joked that perhaps there is some tenuous way that sports gambling is poolable risk to owners, players and coaches but there is real and obvious economic utility with rain. Neither get the advantages of prediction markets.

[JumpCrisscross]

> would argue that a "market" in whether it will rain on Saturday isn't really a prediction market, or even a market, at all. It's just a bookmaking operation

How would you define the difference?

Cat bond premiums absolutely bet on near-term weather odds. I’d argue they’re prediction-esque.

[lxgr]

What would you then call, for example, the markets for weather futures on the CME? https://www.cmegroup.com/markets/weather.html

Are you arguing that weather forecasts (based on either sophisticated modeling or just extrapolating average historical data) are not a thing?

[lxgr]

You seem to be arguing that that's the canonical definition of a prediction market, and anything else, including markets merely aggregating non-insider beliefs about future events, should be called something else. Do you have a better proposal?

[tptacek]

I don't need one.

[lxgr]

Sure, but then discussions about these types of markets are bound to become somewhat confusing for purely nominalistic reasons. (Turns out not every hard naming problem is automatically computer science :)