All writing

Thread · 3 August 2026

Two ways prediction markets break: payout dilution and dead books

A trader can call a market correctly at 10:1 and still lose the edge, because a whale dumping a large position three minutes before settlement dilutes the payout to barely above even. The quieter failure is the market that never gets a book at all, when no market maker will quote it.

Prediction marketsDeFiMarket structureLiquidity

The pitch for prediction markets is that they price the future better than polls or pundits do. That pitch assumes a market which behaves. Two things routinely break the assumption, and both of them are liquidity problems wearing a forecasting costume.

The first is settlement-window dilution. You can be right — call a market two days early at 10:1, get the direction correct — and still watch the payout collapse when a large position arrives minutes before settlement and moves the book against you. Being right about the outcome and being paid for being right turn out to be different questions, and the second one is decided by timing rather than judgement.

The second is the dead book. A niche event can be genuinely interesting and still have no market maker quoting it. With no quotes there is no price to take and no exit if you are already in, so the market exists on paper and not in practice. The event resolves, and nobody who wanted to trade it could.

Both failures point the same direction. The hard part of a prediction market is not the forecast, it is the market structure underneath the forecast: depth, quoting incentives, and settlement mechanics decide whether a correct call is worth anything. Product work on these venues is liquidity design before it is anything else.

Adapted from a thread published 3 August 2026.