How PolyTrace measures Polymarket traders

The threshold applies to one trade, not to a position built up over several. It is a fixed dollar amount rather than a percentile of current activity, which means what qualifies today qualifies next quarter — the bar does not move when the market gets busier, and a quiet week does not promote small trades into the tape.

Trades are read from Polymarket’s live feed and verified on-chain. Sports and esports markets are filtered out; PolyTrace covers politics, crypto, finance, culture, tech, and science.

A position is decided when it resolved one way or the other inside the window. Positions still open in unsettled markets are undecided and excluded. A position that came out exactly even is dropped from the sample rather than counted as a loss — the rate is meant to answer “of the positions that went one way or the other, how many went the trader’s way”.

The window is a fixed 180 days rather than a fixed number of trades. A trade count sounds equivalent but is not: it covers all of history for a wallet with 300 trades and a few months for one with 50,000, so putting the two on one leaderboard compares nothing. That version of the code scored the same wallet 1.91% one way and 13.42% another, depending on which path computed it.

180 days was picked by measuring on production rather than by taste. At 90 days the usable sample collapses; at 365 the results barely move from 180 while far fewer wallets can be measured at all.

Below 20 decided positions there is no win rate. It is withheld, not shown as zero and not shown as a raw fraction. One wallet with 3,606 positions had 12 that could be decided and reported 91.67% — on screen that was indistinguishable from a rate backed by hundreds. A wallet whose history was too long to fetch across the full window is withheld for the same reason: a partial fetch biases the sample toward recent activity.

Sorting a leaderboard on profit alone puts a wallet winning 7% of its positions next to one winning 91%, with nothing on the card to tell them apart. Sorting on win rate is worse: a wallet that won three $5 bets shows 100%. The score combines four measures, weighted:

• Profitability — 35%. Realised and unrealised PnL, logarithmic from $100 (scores 0) to $1M (scores 100), which puts $10k at 50. Losses score 0; how badly a wallet lost is carried by its tag instead.
• Skill — 30%. The win rate after shrinkage, mapped from 40% (scores 0) to 80% (scores 100). Prediction markets pay near even money, so most real wallets land between 45% and 65%; stretching the scale to 100% would compress all of them into the bottom third.
• Breadth — 20%. Distinct markets traded, from 1 to 500. The median wallet trades 33.
• Recency — 15%. Traded today scores 100, silent for 90 days scores 0.

Shrinkage pulls every win rate toward 50% according to how little evidence backs it, by mixing 20 hypothetical coin flips into every wallet’s record. Three wins from three positions reads as 56.5%; 500 from 500 still reads as 98%. The extreme has to be earned. This applies to the score — the win rate shown on a trader page is the raw one, next to its sample size.

All four scales are absolute, not percentile, so a wallet scores the same today and next quarter regardless of what other wallets do. A wallet that cannot be given a trustworthy win rate is unrated rather than scored low: ranking an unknown below a demonstrably bad wallet is a claim the data does not support. About two thirds of profiles are unrated on this basis.

• No odds weighting. Shrinkage answers how much evidence backs a rate, not how hard the bets were. A wallet that bought twenty-five 97-cent near-certainties and won them all scores like one that won 65% of four hundred coin flips. Statistically that is defensible — twenty-five straight wins happen by chance about three times in a hundred million — but it is not what a follower wants to know: the first wallet collects pennies at low risk, the second is actually forecasting. Separating them needs each position’s entry price, which the profile does not carry. It is a data-model gap, left open rather than papered over with a heuristic.
• Win rate is the wrong question for some wallets. Market-making and arbitrage strategies take many tiny losses by design. A low win rate on such a wallet describes its strategy, not its skill.
• Profiles key on proxy wallets. A Polymarket proxy wallet is the address Polymarket exposes publicly, which is not necessarily the owner’s main wallet. One person may hold several; PolyTrace does not link them.
• PnL includes open positions. Unrealised profit moves with the market and is not money taken off the table.