Guide

Does copying Polymarket whales work?

We re-priced thousands of graded whale bets as if you copied them 1 minute to 24 hours late. Being late barely matters. Picking the wallet is everything.

The usual worry

The common complaint about copy trading is speed: by the time you see a whale's trade, the price has moved. We tested that directly by pricing each bet again 1, 5 and 15 minutes, and 1, 6 and 24 hours after the whale's fill.

What the data shows

As of October 2026, the average price move against a late copier is small: well under 1¢ at 1 to 15 minutes, and around 1 to 2¢ after 6 to 24 hours, in sports and politics alike. In thin markets the quoted price can be far from any real fill, so we leave those out.

The real problem

Across all graded $10k+ trades, whales lose to the price by about 3 to 4 points. Copying a random whale on time just copies a losing bet on time. The question is which wallets, if any, beat the price on bets they placed after we started watching them.

The honest test

We pick wallets using only bets that resolved before a cutoff date, then measure them only on bets placed after it. That out-of-sample number is shown first on the findings page, with its error bar, even when it's too small to call.

Before you copy anything

Check the sample size, the categories a wallet trades, and whether its record holds without its single best bet. Nothing here is a recommendation; it's a record of what happened.

Live right now

5759 of 10564 graded whale flags called vs 6144.7 expected from price: -3.7 pts. Scoreboard →

More guides

Share on X

Data only. Not financial advice.