A 46,945-wallet sample still needs careful questions

A dataset with 46,945 wallets sounds authoritative. It is certainly more interesting than one winning screenshot. The number alone still cannot tell you what was measured.

Treat that figure as the size of a reported sample to investigate, not proof of a particular profitability result.

Reconstruct the accounting

Does profit include open positions? Are fees included? How are transfers and linked wallets handled? Does the sample cover one contract type or the whole platform?

An address that bought before the observation window may appear to sell profitably inside it while its original cost is missing.

Compare roles carefully

A maker posts a resting quote; a taker trades against available quotes. Different results between those groups can reflect skill, selection, market conditions, or the kinds of trades they choose.

A group average does not establish that switching order type would give a particular trader the same return.

Look beyond the headline

Ask for the median result, loss distribution, concentration among top wallets, and sensitivity to the chosen period. A small number of extreme outcomes can dominate an average.

Try to reproduce a narrow slice with the stated method. If the definitions are unclear, the sample is difficult to audit regardless of its size.

Big data is useful when it makes the uncertainty smaller and more explicit. It should not make the questions disappear.

Further reading: Prices and order books · Resolution rules · Current fees.

Explore the markets: Open Polymarket (referral link).