Polymarket One-Sided Books

A one-sided snapshot has liquidity on only the executable side — resting orders at the 0.99 bound with no counterpart on the other side. It is structural on this venue: outcome markets concentrate depth on the likely winner's bound as resolution approaches.

The measured distribution is stark: 16.9% of 5-minute snapshots are one-sided overall, 76.2% of final-minute snapshots have an empty bid or ask side, and 90% of final-second books carry a single side.

The mechanism is settlement pressure: as a reference print resolves the contract, the losing side's liquidity exits (or never enters), and the winning side rests deep at the bound — measured at a median 42,254 shares.

Between the decision and the print, one-sidedness is the venue's operational texture: a trader who wants the other side must limit into the void or cross the venue entirely.

Against a one-sided book, your executable choice is the resting side at its price (join the bound) or a crossing order that finds no queue (you become the book). Both appear in the record as depth changes and tape prints.

The loaded performance habit is measuring your fill against the executable side rather than an imaginary mid — the mid's honest definition includes the one-sided case.

One-sidedness is a feature to study, not a defect to filter: the archive publishes the one-sided flag on every row and the fill-simulation pages use it, because hiding the state would fabricate liquidity.

• Crossed books
• Illiquidity
• Imbalance data