Whale Following on Polymarket: Copy Large Wallets Without Copying Their Mistakes
How to follow large Polymarket wallets intelligently: reading position context and entry prices, spotting hedges and market-making inventory, sizing a copy, planning exits and knowing when a whale is not a signal.

Polymarket is transparent by design. Trades settle onchain, positions are visible, and leaderboards rank wallets by profit. That transparency makes whale watching feel like an edge: if a large wallet is buying, the reasoning goes, the wallet knows something. Sometimes it does. Often it is hedging, market making, or closing a position in pieces.
Copying a whale is not copying a trade. It is copying a position at a different price, with a different bankroll, a different time horizon and no access to the whale's reason. The work is to reconstruct the context before you copy anything.
Read the position, not the trade
A position is the result of many decisions. Start with the average entry price, the current price, the size, and the unrealized profit or loss. Then ask what fraction of the wallet the position represents. A $200,000 position in a $2 million wallet is 10%; the same position in a $200 million wallet is a rounding error.
Look at the history. Was the position built in one trade or twenty? Did the wallet add on the way up or average down? Did it sell any of the position, and at what price? A wallet that bought at 0.30 and is still holding at 0.55 has a very different risk profile from one that bought at 0.54 yesterday.
Check the market's liquidity. A $200,000 position in a market with $50,000 of daily volume is a large share of the book. The whale cannot exit without moving the price, which means the whale's exit is your risk.
Entry price versus current price
The most common copy-trading error is treating a whale's entry as a signal at today's price. It is not. The whale's edge was the difference between the entry price and the value at the time. Once the price moves, that edge is partly realized and partly paid away.
Suppose a whale bought 100,000 shares at 0.30. The market is now 0.55, and you copy at 0.55. The whale is up 25 cents per share. You are starting at the current consensus. If the true probability is 0.58, the whale's remaining edge from here is 3 cents before fees, and your edge is the same 3 cents, but you paid 25 cents more for it. The whale can absorb a 20-cent drawdown and still be profitable; you cannot.
The question is not "what did the whale buy?" It is "what is the remaining edge from the current price?" If you cannot answer that independently, you are not copying a trade. You are providing exit liquidity.
Hedging and market-making activity
Large wallets are often not directional. A market maker quotes both sides and ends up with inventory that looks like a position. A hedger may hold Yes in one market and No in a related market, so the visible leg is not the view. A negative-risk conversion can turn a No position into Yes positions across an event, which looks like a big bet but is actually a relative-value trade.
Check the related markets before you interpret a position. If a wallet is long Yes on one candidate and long No on another in the same exclusive event, the net exposure may be small. If the wallet has a long history of two-sided fills, it is probably providing liquidity, and the inventory is a cost of doing business rather than a forecast.
Polymarket's maker rebates and liquidity rewards give large wallets a reason to quote even when they have no view. A wallet earning rewards is not necessarily expressing an opinion about the outcome.
Size the copy, not the whale
Copy in percentage terms, not dollar terms. If the whale risks 2% of a $10 million bankroll, that is $200,000. Your equivalent is 2% of your bankroll, or $200 on $10,000. Copying the dollar amount is impossible, and copying the share count is worse, because it ignores the price.
Then adjust for correlation. If the whale holds five positions in the same election, the whale's true exposure is the portfolio, not the single market. Your copy should respect the same concentration limit, or you will be more concentrated than the person you are copying.
Finally, adjust for liquidity. If your order is large relative to the book, you will pay more than the whale did. Compute the fill price for your size, not the midpoint.
Exits are half the trade
Whales do not announce exits. They sell in pieces, often into strength, sometimes through the same book you are watching. A copy that only mirrors entries has no exit rule and will hold through the reversal.
Set your exit before you enter. It can be a price target, a time stop, or a change in the underlying facts. It can also be "exit when the whale's position stops growing", but that is a weak rule because the whale may be patient while you are not.
Watch for distribution. If the wallet's position is shrinking while the price is flat, someone is selling into the bid. If the price is rising and the position is shrinking, the whale is taking profit. Neither is automatically a reason to sell, but both are information.
When a whale is not a signal
A large position can be noise. New wallets can be funded by the same person and used to split orders. Test trades can be large in dollar terms but small relative to the wallet. Airdrop farming and volume incentives can produce activity with no view. Wash trading between related wallets can create the appearance of conviction.
Sports markets add another layer. A whale may be hedging a sportsbook position, or arbitraging a bonus, or managing a portfolio of correlated games. The position is real, but the motive is not a forecast of the game.
Reputational risk also matters. Copying a wallet that turns out to be trading on non-public information can expose you to scrutiny even if you did nothing wrong. If a position looks too well timed, assume you do not know why, and size accordingly.
How to execute
- Build a watchlist of wallets with a public history longer than a few months.
- Log the entry price, size, and the price at the moment you see the trade.
- Compute the remaining edge from the current price, independently.
- Check related markets for hedges and net exposure.
- Check whether the wallet is a market maker or a rewards farmer.
- Size as a percentage of bankroll, then reduce for correlation and liquidity.
- Predefine the exit and the maximum loss.
- Recheck the thesis if the whale adds or reduces.
Risk controls
The main risks are adverse selection, crowding, liquidity and model risk. Adverse selection: you see the trade after it happened, so you buy from someone who may know more. Crowding: many copycats chase the same wallet, which pushes the price and makes the exit worse. Liquidity: the whale's position is larger than the book. Model risk: you are inferring intent from a position.
Cap any single copied position at 0.5% to 1% of bankroll, and cap the total copied exposure at 5%. Keep a log of every copy with the whale's entry, your entry, and the outcome. After twenty trades, you will know whether you are copying skill or copying noise.
Checklist
- The wallet has a track record longer than the position you are copying.
- You know the average entry price and the current price.
- You have computed the remaining edge yourself.
- You have checked related markets for hedges.
- The wallet is not primarily a market maker or rewards farmer.
- Your size is a percentage of your bankroll, not the whale's dollar amount.
- Your exit is written down before you enter.
- The market can absorb your exit without a double-digit price move.
Whale watching is a research tool, not a strategy. The position tells you that someone with capital made a decision. It does not tell you why, what they paid, what they will do next, or whether the edge is still there. Copy the analysis, not the ticket.
Sources
- 1.Data API Overview— Polymarket Documentation
- 2.Positions and Tokens— Polymarket Documentation
- 3.Trader Leaderboard— Polymarket Documentation
- 4.Public Analytics— Polymarket Documentation
- 5.Fees— Polymarket Documentation
- 6.Interpreting Prediction Market Prices as Probabilities— National Bureau of Economic Research