2205

LESSWRONG
LW

2204
Prediction MarketsWorld Modeling
Frontpage

7

Prediction markets for social deduction games without weird incentives

by Mikhail Samin
15th Nov 2025
2 min read
0

7

This is a linkpost for https://mikhailsamin.substack.com/p/game-markets
Prediction MarketsWorld Modeling
Frontpage

7

New Comment
Moderation Log
More from Mikhail Samin
View more
Curated and popular this week
0Comments

Being able to N/A contracts with market manipulators reduces the incentives to do bad stuff.

Prediction markets are fun! Social deduction games are fun! Can you merge the two?

The simple solution is normal games plus betting among the spectators.

But for a long time, I wanted to play a social deduction game with prediction markets that players can participate in, however, I didn’t really try to figure out how to make it work.

Straightforward combination of prediction markets with social deduction games distorts incentives. Either the bad guys might want to sell their probability of winning for providing everyone information, or they might want to increase their chance of winning by losing bets to mess up prediction markets.

Recently, I thought of a solution.

Very simply: at the end of the game, cancel all bets that the bad guys participated in. (Or, in some games, bets between sides.)

Additionally:

  • Allow orders that restrict the set of possible counterparties (e.g., I want only Alice and Bob to be able to trade with me on an order).
  • Allow orders with limits per counterparty (e.g., anyone can buy at most 3 No shares).
  • Have displays for market price history for trades between selected sets of participants (e.g., I might want to be able to see the price history of trades only between Alice, Bob, and Carol, but not trades with Mallory; or all trades that Mallory made).

Since none of the bets the bad guys make get executed, they cannot lose or gain market currency by trading, which removes the weird incentives.

So: now markets simply facilitate and make clearer public information exchange and everyone’s collective beliefs.

The bad guys might want to participate in the markets the same way they would want to participate in the public discourse: to mislead others about their roles and the situation, to try to pretend to be on the good side and play the way someone on the good side would.

Since groups of players can exclude any particular players from significantly influencing the in-group trading, the bad side can only distort information to the extent it would be able to do so by talking.

I expect that for a game between good players familiar with and enjoying prediction markets, it might be the same as regular games, except a lot more fun.

It likely won’t be as chaotic as social deduction games with prediction markets that allow market manipulation, as it’s much clearer what to do: if you’re a bad guy, simply pretend to be a good guy as well as you can.

I would be excited about someone implementing the markets with selections for sets of players who are the counterparties or are displayed.

It could be implemented as a whiteboard where people write down bets and orders (increases friction and makes tracking what’s going on harder), as an app (decreases friction but requires looking away from the people to look at the screens), or as some combination of various input and output mechanisms.

(It would also be interesting to compare the results and histories of trading between the players on the good side to trading between the spectators.)