I am a decently experienced forecaster and trader on prediction markets, and often partake in efforts to grow the community to improve forecasts. For those not in the loop, check out this video.

I often notice that new predictors and traders, who often have little capital, trade highly suboptimally. One such example is locking up capital more than necessary by not selling shares when a market is almost certain to resolve in one direction.

General limit sell strategy

Buy some amount of YES shares, bought at a price quite a bit below the expected value of the shares. For Manifold, the Manifolio calculator automatically Kelly bets for you. After, place a NO limit order at a higher price. Wait for the market to correct upwards, and wait until your limit sell order is complete.

Avoid rules lawyering and misresolutions

For decentralized markets such as Manifold and Polymarket, there are some weird quirks that occur when resolving a market YES or NO. On Manifold, the creator of a question resolves the market, or if the creator is inactive, a moderator takes over resolution. This can take a day to even longer for moderators to verify if the resolution criteria is fulfilled. Polymarket’s approach is far worse, as disputes are resolved by the centralized colluding cabal known as UMA, which the professional trader Domer is probably criticizing in his username right now.

Even on the centralized exchanges, such as PredictIt and Kalshi, there can still be ambiguous resolution criteria or information a trader may have missed. The most famous example was back in 2018, Rick Scott decided to serve out his governor term, taking his senate seat days after the senate control market was set to resolve.

The simplest method to avoid this risk is to sell before the madness!

Efficient capital allocation

Clearly, the miniscule difference (after possible fees) for not waiting is a small cost to pay, compared to the alternatives. Every uninvested unit of currency is an opportunity cost. For Manifold, use this calculator to determine your personal risk-free interest rate. This is your opportunity cost of not selling.

Per Isaac, the creator of the : “For example, the LK-99 market is currently at 3.12% and resolves in 13 months, so if we assume that the correct probability of it being a superconductor is 0.01%, betting NO in that market is equivalent to a 3.28% monthly compounding interest rate. As such no one would ever want to issue a loan for lower interest than that.”

For real world markets, your invested funds do not earn interest. This is even a better incentive to limit sell.