Vitalik Buterin pfp
Vitalik Buterin
@vitalik.eth
Prediction markets truly are a super valuable tool for making sense of events in real time. https://polymarket.com/event/sam-back-as-ceo-of-openai
11 replies
39 recasts
193 reactions
tldr (tim reilly) pfp
tldr (tim reilly)
@tldr
Could prediction markets incentivize anti-expected behavior? Eg: 1. Prediction Market launched about whether Person A will take Action X 2. “No Action X” is signaled (by others and possibly by Person A) 3. Person A loads on “Yes Action X” shares at very low price 4. Person A takes Action X & profits
1 reply
0 recast
0 reaction
Jack Miller pfp
Jack Miller
@jackm
That's kind of the point - step #3 moves the price and price movement and/or whale accumulation acts as a signal The extreme case where Person A commits fraud to intentionally manipulate the market is not itself without risk and would only work once at most
1 reply
0 recast
0 reaction
tldr (tim reilly) pfp
tldr (tim reilly)
@tldr
Why once only? They just need to create anonymous shells to do the purchasing. Can also distribute the buying (so doesn't have to appear as a whale) Lots of people already take similar actions with tokens and NFTs and we have no idea who they are
2 replies
0 recast
0 reaction
kugusha 🦋 pfp
kugusha 🦋
@kugusha.eth
this sounds like a sybil resistance problem. 1 extra layer of web3-native verification should be good enough to cover most cases of this behaviour
1 reply
0 recast
0 reaction
tldr (tim reilly) pfp
tldr (tim reilly)
@tldr
Reputation lengthens the game, which is good But even still, the target of a prediction *by necessity* always has both (1) asymmetric information (2) short term financial incentive to act counter to the market This imo will always be a potential problem for keeping predictions predictive rather than pervertive
1 reply
0 recast
0 reaction
Jack Miller pfp
Jack Miller
@jackm
Bringing asymmetric information to light is the point. The case where someone can directly and unilaterally manipulate the outcome of the market is an edge case and not a very good market to play in. If those markets are unfairly manipulated, people will stop participating.
1 reply
0 recast
0 reaction
tldr (tim reilly) pfp
tldr (tim reilly)
@tldr
Advantaging the asymmetric information of the actors themselves in a way that affects the outcome is certainly *not* the point. And I agree this could ultimately cause people to stop participating, which is why it's a potential limitation of PMs. As to whether it's an edge case, check out Polymarket homepage:
1 reply
0 recast
0 reaction
Jack Miller pfp
Jack Miller
@jackm
In a world of perfect information symmetry there is no point to a prediction market If Sam Altman mass bought "Sam Altman announces new company" shares then the price would go up and we'd have a signal Then ppl would see one guy can just decide the outcome and they'd bet on i.e. the "Democratic Nominee 2024" instead
0 reply
0 recast
0 reaction