Win Rate vs Payoff: What 9,883 Copied Trades Say

A win rate counts how often a trader was right and leaves out what each win paid. On Polymarket the second number depends on the price of the share, so two traders with the same win rate can end far apart. In our replay of 9,883 copied trades, 58% ended in profit, yet the average copy made only about $1.50 per $100.

Key numbers

• 58% of 9,883 copied trades ended in profit, and the average $100 copy made about $1.50.
• Before fees, the win rate needed to break even equals the price paid: 80% at 80¢, 30% at 30¢.
• The wallet with 91% of copies in profit made $8.70 per $100 copy. The one with 66% made $34.40.
• In one week of settled positions, shares bought at 90¢ to 98¢ ended in profit 89.7% of the time and returned −0.9%.

Can you win most trades and still lose?

Yes. In one week of settled positions, shares bought at 90¢ to 98¢ ended in profit 89.7% of the time and returned −0.9%. In our 120-day replay, 58% of copies ended in profit, yet the average copy made only about $1.50 per $100.

We took the 166 wallets that passed our skill filter on 30 September 2026 and replayed every position they closed in the 120 days before that date, 85,408 in all. Three trade rules kept 9,883 of them. Each kept position was copied with $100 and closed when the whale closed.

The typical copy did well: the median made about $18 per $100, and 58% of copies ended in profit. The average copy made far less, about $1.50 per $100, because losing copies cost more on average than winning ones earned. Most of the total came from a handful of wallets. The full study is in what we learned from 15,371 wallets.

A positive median next to a flat total looks like a contradiction until you look at what a win and a loss are each worth.

Why does the entry price decide the payoff?

A winning share pays $1. The price you pay sets what a win earns and how often you must win to break even.

Every Polymarket share is priced between $0 and $1 and pays $1 if its outcome happens. Buy at 80¢ and a win adds 20¢, while a loss can take the whole 80¢. For $100 put in at each price:

The last column is the price itself. A wallet that buys at 90¢ and wins 92% of the time is two points ahead of break-even. A wallet that buys at 30¢ and wins 40% of the time is ten points ahead, with a win rate most people would call poor.

The loss column says "up to" because a trader can sell before the market resolves and keep part of the stake. Selling early also means a win can earn less than the table shows. Trading fees are left out too: Polymarket's taker fee, charged in most market categories, moves each break-even figure up by two points at most.

What did the highest win rate earn?

It earned the least per copy. Among our four strongest wallets, the one with 91% of copies in profit made $8.70 per $100, and the one with 66% made $34.40.

These four were the strongest of the 166 wallets over the 120 days, each copied at $100 per trade with the three rules.

Whale D was in profit on nine copies in ten and earned the least on each, because its wins were small. Whale A ended a third of its copies without a profit and made about four times as much per copy.

Whale C shows a third way to make money. Its average per copy was close to Whale D's, but it was copied 583 times, and its total was the largest of the four at +$5,296.

They were picked after the fact, by their result in the same window they are measured on, so treat them as illustrations.

What does one week of settled positions show?

The two bands with the highest win rates lost money, and the band with the lowest win rate returned the most.

Our report for week 41 covers 99,946 positions settled between 4 and 11 October 2026 by wallets in our 1,049-wallet directory.

Positions bought at 90¢ or more ended in profit almost nine times in ten and lost money in both bands. Positions bought under 10¢ ended in profit about one time in seven and returned 18.4%, carried by a few long shots.

One week is a small sample and a single resolved market can move a whole band, so the numbers will differ next week. What holds every week is the arithmetic: the higher the price, the more often a position must win to break even.

What should you read instead of win rate?

Win rate is still worth a glance once you know the prices behind it. Four other numbers say more about a wallet:

• Return on the money it put in. It already combines how often the wallet wins with what each win pays.
• The deepest drawdown, which shows what following the wallet would have felt like at its worst.
• The share of profit that came from one trade. A wallet whose record rests on a single position has shown very little.
• The number of settled trades. A return measured on ten trades is mostly noise.

These are our four default checks: a return of 15% or more, at least 30 settled trades, a drawdown of 25% or less, and no single trade that made more than half the profit. The methodology page describes each one. Every page in our whale directory shows a wallet's win rate next to its return and its settled trades, so you can read them together.

What are the limits of these numbers?

The replay is a model. It puts $100 into each position that passes the rules and exits when the whale exits. It does not model slippage, a separate fee charge, a stop-loss or a take-profit.

The four wallets in the table were chosen by their result in the window they are measured on, so their figures flatter them. The weekly table covers seven days, and past results do not predict future ones.