Paste a Polymarket wallet. See whether its picks beat the prices it paid.
A profit board shows who made the most. It does not show whether that wallet bought at prices that were too low for how things turned out, more often than chance allows. This page reads every resolved position and answers that.
Hyperliquid wallets: the wallet check.
How it reads a wallet
One question per position. A share bought at price p pays $1 if it wins. With no skill, it wins with probability p, so the edge, outcome minus price paid, averages zero. Someone who keeps buying at prices that turn out too low has a positive edge, and z measures how far that is from chance.
Bets on one event are one bet. Ten positions on the same election share one outcome. z sums the edge within each event before comparing it with chance, so a wallet cannot look strong by splitting one view into many positions.
Buying favourites is not skill until one loses. A wallet that only buys at 0.95 and up wins almost every bet, by about five cents each. Its edges are tiny and all positive until the first loss, so a spread measured from those edges alone is tiny too and z looks huge: thirty wins bought at 0.995 scored 5.48 before 2026-10-18. z's spread now never falls below what chance alone would give (p(1 − p) per one-sided market), and those thirty wins score 0.39.
Volume is not scored. Trading volume can be made by trading with yourself: WIRED found over $600 million traded on two Polymarket election markets after the results were known, which one analyst put at up to 99.6% wash trading in some weeks (WIRED, 2026-10-15). z never reads volume: each market counts once at its average buy price, however many times it was traded, and a price that matched the odds has no edge to add. Buying a winner after the result is known adds a few cents per market, and the floor above keeps that from counting as skill.
The losers Polymarket lists elsewhere. A losing share is worth nothing, so many are never redeemed. They stay among open positions, not closed ones. Counting only closed positions made 11 of the profit top 50 look like stars in audit 008; this page counts both.
Two bars. z ≥ 2 is the usual bar for a wallet you came to with no prior selection. A wallet found on the profit board was picked from at least 450,853 wallets, so it needs z ≥ 5.33. That is a floor: the board reports no total, and a username search on it returns a wallet ranked 450,853. Of the top 50 in audit 008, none came close; the highest z was 2.40.
What z cannot see. Position sizes (each position counts once; a few huge winning bets can make the money while the picks overall are ordinary), and whether the wallet traded on information others did not have.
Not investment advice. Public data from data-api.polymarket.com, fetched by your browser; the wallet is never sent to us, and each check adds one tally with no address. Method: audit 008, tools/audit_polymarket.py.