How we pick Polymarket traders to follow

What we measure

The window is the last 90 days and only resolved markets count. For each wallet we show win rate, ROI (realized plus unrealized over capital invested), max drawdown, resolved markets, average entry price, median trade size and median holding time.

Wallets with too few trades are too small a sample and get no detail page.

Exclude first, then rank

One-off bets: wallets whose results mostly come from a single event are excluded; too much of that is luck.

Wash volume and self-crossing: wallets whose volume dwarfs their results, or that trade against themselves, are excluded.

Market makers: wallets that quote both sides earn the spread, which copying their direction can't replicate, so they are excluded.

Wallets that pass are ranked by long-term results. The composite used for ranking is internal and not published.

Sources and updates

Fills come from Polygon on-chain records and Polymarket's public APIs. Website pages are rebuilt daily and show their data time in Beijing time. Values we can't get show "No data" and are never filled with 0.

Limitations

Past 90-day results don't predict the future. The filters remove obvious noise but not all luck, and they can't see whether the person behind a wallet has changed strategy.