Who actually wins on Polymarket, and which markets look off.

Market trades 10¢ on YES (deeply against the outcome) on 1360K in lifetime volume. Tail-priced markets with active volume are worth a second look against the underlying question and any new information.

Market trades 10¢ on YES (deeply against the outcome) on 2352K in lifetime volume. Tail-priced markets with active volume are worth a second look against the underlying question and any new information.

Market trades 10¢ on YES (deeply against the outcome) on 109K in lifetime volume. Tail-priced markets with active volume are worth a second look against the underlying question and any new information.

Market trades 1¢ on YES (deeply against the outcome) on 43894K in lifetime volume. Tail-priced markets with active volume are worth a second look against the underlying question and any new information.

Market trades 1¢ on YES (deeply against the outcome) on 32524K in lifetime volume. Tail-priced markets with active volume are worth a second look against the underlying question and any new information.

A plain answer to where Polymarket payout money comes from: not the platform, but the traders on the losing side of the same market, whose collateral is locked in the Conditional Token Framework contract from the moment a share pair is created. Covers the split, merge and redeem operations that make the accounting close, the check that found all 400 markets in the July 31, 2026 snapshot pricing YES plus NO at exactly 1.00, a side-by-side of the peer-funded structure against a bookmaker taking the other side itself, the oracle step that has to finish before collateral is released, and what the arrangement does and does not protect against. Editorial, not advice.

Polymarket's past odds are free to read through the market page chart and through a public CLOB endpoint that needs no key, but the depth is far shallower than the parameter names imply. Covers the four routes to a past price series and what each costs in effort, the prices-history endpoint and its token-id-per-outcome quirk, ten requests we sent on August 1, 2026 with the point counts and spacing that came back, the minimum-fidelity floors that make a bare interval=1m return HTTP 400, and the finding that interval=all and interval=max both returned about 31 days starting July 1, 2026 on three separate markets including two open since 2025. Also why a thin market's price spike can represent very little money, and why this site publishes a dated snapshot rather than a time series. Editorial, not advice.

Polymarket's original market has always run on Polygon PoS, chain ID 137, and its developer documentation lists Polygon mainnet as the only network its contracts occupy. Covers the five core contracts and their current addresses after the April 28, 2026 migration to CTF Exchange V2 and pUSD collateral, the hybrid design that keeps order matching and cancellation off-chain while settlement, custody and resolution stay on-chain, why an unfilled order leaves no public trace, how the Conditional Token Framework turns collateral into ERC-1155 outcome tokens priced between one and ninety-nine cents, and why the separate CFTC-regulated Polymarket US exchange publishes no ledger at all. Editorial, not advice.

Polymarket publishes no official volume statistic, so every figure in circulation is a third-party reconstruction, and the reconstructions disagree by up to two and a half times on the same month. Covers the monthly 2026 series side by side, the three differences that explain the spread (which venue is in scope, whether each fill is counted once or twice, and how much raw flow is filtered), what we read from the public API on July 30, 2026 across 2,091 open order-book markets, why the busiest single market held 18.1 percent of a day's volume while the median market traded $4,107, and why summing today's open markets understates the platform by roughly ten times. Editorial, not advice.