84 % of Polymarket Wallets Lose: How to Read Other Traders' Trades

Every Polymarket trade runs through smart contracts on Polygon, so other people's entries, exits and sizes are public. A whole industry of wallet trackers and trader rankings grew out of that. They stay useful exactly as long as you remember one thing: a wallet shows the action and hides the reason.

What is visible

A trading account is a smart contract wallet on Polygon. Every buy and sell lands in the chain with the address, the market, the side, the number of shares, the price and the time.

Trackers read those records, group them by address and total up a result. No private access is involved, which is why the data is plentiful and the rankings are everywhere.

The same holds in reverse. Your trades are visible to everyone else on the same terms.

Five signals a wallet gives

Entry time against the news. You can see whether an address bought before or after a publication. This is the strongest available signal, because it needs no guess about motive.

Entry price. You can see where the position was built and how far that sat from the current price. Entry at 12¢ in a market now trading at 60¢ tells you when the decision was made.

Size against the address's own history. A $50,000 trade from an address that usually trades a thousand is an event. The same trade from an address turning over millions is routine.

Side and any switch of side. You can see whether the address added to a position or flipped it. A flip in the middle of an event says more than the original entry.

Sequence of actions. You can see whether the trader scaled out in pieces, sat until settlement or left on the first move. That is temperament, and it repeats across trades.

Three traps

Motive stays invisible. A $50,000 purchase of No can be a directional bet or the closing of risk on a position this wallet never held. A hedge and a conviction look identical on chain.

Part of the picture sits outside the wallet. A trader can run several addresses, hold positions elsewhere and carry obligations the chain knows nothing about. You see one slice of a decision made on the whole.

One wallet is not one person. An address can be a fund, a group, money under management or a bot executing somebody's strategy. Top trader rankings pile all of that into a single list.

Why copying after the fact breaks

Time passes between somebody's trade and your reaction, and the price moves inside it.

Run the arithmetic. An address bought at 30¢, you saw it ten minutes later with the market at 38¢. Yours is a different trade: a different balance of risk and payout, less room to settlement. Add the spread you pay entering and leaving.

So another trader's entry works as a reason to look at a market rather than as an instruction. By the time you see it, whatever edge existed has been partly paid into the price.

Counting somebody else's result honestly

Rankings measure in different ways, and the method changes the number.

Realized and unrealized profit are not the same thing. An address deep in the green on open positions can close the year negative. The observation window decides as much: a result across three good months and a result across two years are different claims.

Then the survivorship correction. You are looking at the ones still trading. Addresses that went to zero and stopped do not appear in any top list, though they may have started with the same approach.

What to do with trader rankings

A ranking answers the question of who made money. Trading asks what is underpriced now. The first does not reduce to the second.

Two uses hold up. A ranking gives you a list of addresses whose full history is worth reading, rather than one bright trade. And it shows which markets attract large money, which tells you something about liquidity.

What it does not give is reasons. An address at the top might have held one correct position for two years or made four hundred trades at a thin edge. Those are different skills, and the second cannot be copied by mirroring the next trade.

How many traders are losing

One number belongs in your head while you read any ranking. In an on-chain analysis of 2.5 million Polymarket wallets published in April 2026, researcher Andrey Sergeenkov found 84.1 % of addresses unprofitable. About 2 % of addresses had accumulated more than $1,000 in profit, and 840 addresses, or 0.033 %, more than $100,000.

Other counts land elsewhere. Predict Parity's on-chain data puts 69 % of traders at a loss, and a December 2025 study by the analyst DeFi Oasis, covering 1.7 million addresses and 124 million trades, found 70 % of addresses unprofitable. The spread between those numbers is method: counting addresses or trades, including open positions or not, and the window chosen.

The range from 69 to 84 % says the same thing either way. Profit concentrates in a minority, and the top of a ranking is the tail of a distribution rather than a template.

Insider cases: seen and inferred

An entry hours before the news is visible on chain, and that is a fact. Knowledge of the news is not visible on chain, and that is an inference.

The line matters. You can say an address bought Yes at 8¢ four hours before an announcement, and the price stood at 70¢ when it landed. You cannot say how the person knew unless a named source says so.

For a reader the value of such a case is the method rather than the verdict. It shows how a price moves around an event and at which point the market repriced the outcome.

Turning somebody's trade into your own hypothesis

Take the trade and reconstruct the question the person was answering. State your own probability for that question before you look at the current price. Then look at the price and measure the gap. Check the spread and the depth: is the gap large enough to survive entry and exit. Decide on your own arithmetic rather than theirs.

Used this way, another trader's entry is an attention filter. It cuts thousands of markets down to one worth calculating.

Your wallet is visible too

The symmetry runs both directions. Your entries, sizes and exits are available to anyone reading the chain.

Two consequences follow. A large order in a thin market shows up in the book and in the chain. And the trade history you review in the terminal looks to an outsider exactly like somebody else's looks to you: action without reason.

Six questions, and the last one is about you

Three cover the entry: when it landed relative to the news, at what price, and how the size compares with this address's usual trades.

Two cover what came after: whether they added or flipped, and whether the position is closed or still running.

The sixth is yours. What question was this person answering, and do you have an answer of your own? Without one, you have read a story rather than a market.

Your next step

Reading other wallets pays off when you read your own the same way. Open the Pulse terminal, go to Analytics and look at your history with the same eyes: entry times, prices, sizes, side. The Price vs. win block shows the one thing another wallet never will, which is how well your probability estimates match your results.