How Polymarket Crypto Markets Settle: Price to Beat, Reference Prices, and Your Data

Binary outcomes, priced 0 to 1

Every Polymarket crypto up/down market has two outcomes — Up and Down — each quoted between 0 and 1. The price is the market's implied probability: an Up token at 0.62 is the book pricing a 62% chance the asset finishes up over the window. At settlement one outcome is worth $1.00 and the other $0.00.

Because the market is binary, the two outcomes are exact complements: the Down price is 1 − the Up price, with the resting sizes preserved. You only need one side's book to reconstruct both — but you need the book, not a single last price, to see the spread and the size a real fill would have paid.

The price to beat

An up/down market needs an authoritative number for where the underlying sat at the open and where it lands at the close. Polymarket sets a 'price to beat' at the market's open and compares it to a final reference price at the close; the open and close references come from the market's own settlement anchors (Chainlink-published reference prices), not any single exchange's last trade.

A real settled example from our archive: the price to beat was $62,701.75 at the open, and the final reference printed $62,519.65 at the close — below the line — so the market resolved Down, and the Up token's price slid toward 0 across the window as spot fell. The precise reference source and the settlement timing are defined per market in Polymarket's own rules and can differ; read them there before trusting a backtest's settlement logic.

What the settlement record actually looks like

Resolution fields — the price to beat, the final price, the winner, and the volume — backfill roughly 12 to 24 hours after a market settles. Across the 60,307 settled crypto up/down markets DepthFeed has captured open-and-close for, 49.2% resolved Up and 50.8% Down — a near coin-flip you can only confirm by measuring the real settlement record rather than assuming it.

That near-50/50 split is exactly what you would expect from an efficiently priced short-dated market, and it is a useful sanity check: any backtest whose settlement assumptions skew far from it is probably mislabeling winners.

Why settlement timing matters for your data

If you backtest a strategy that holds into settlement, two things decide whether the result is real: the book right before the close (where spreads widen and depth thins, so fills get worse) and the underlying reference at the settlement instant (which determines the payoff).

DepthFeed addresses both. Event-driven CLOB capture means you have the book as it actually was into the close — every re-quote, not an hourly sample — and every snapshot joins to a high-frequency underlying reference price by epoch-millis timestamp, so you can line the book up against the move that determined settlement.

The data you want around settlement

• The full bid/ask book into the final minutes — spreads widen and depth thins as the close approaches.
• Epoch-millis timestamps, so you can align book state to the settlement instant.
• The price to beat and the final reference, to label the winner exactly as the market did.
• Event-driven capture, not hourly — the last minutes of a 5- to 60-minute market are where settlement risk lives.