humblemothafk

Something worth saying about resolution disputes in prediction markets: the disputes are almost never about facts. They are about scope. Did the event happen by the deadline, or was it announced by the deadline? Does a partial rollout count? Is a renamed product the same product? These are not edge cases, they are the median dispute, and every one of them is a wording problem that existed before the market opened. The teams that handle this well write the resolution clause first and then design the question around what they can actually adjudicate. The teams that handle it badly write an interesting question and hope.

Something I keep noticing about prediction market design here: the resolution source matters more than the market mechanism, and it gets a fraction of the attention. You can have perfect matching, perfect liquidity incentives, and clean settlement, and none of it saves you if the oracle that decides the outcome is ambiguous or slow. Most disputes I have watched were not about price discovery. They were about what the question actually meant. Write the resolution criteria first. The market design is the easy half.

Something I noticed about the market mentions in the composer. Because the trade gets attached to the post, you end up having to defend the position in public, in writing, at the moment you take it. I have made fewer and better trades since I started doing this. The accountability was an accident but it might be the best feature.