Does Polymarket copy trading work? The honest math

Copy trading copies a trader's decisions, not their returns. A worked example with Polymarket's real fee formula, which traders copy well, which don't, and how to size so the edge survives.

polymarketcopy tradingstrategyfees

"Does copy trading Polymarket work?" is one of the most common questions about prediction-market bots, and most answers online are either sales pitches or blanket dismissals. The truthful answer is narrower: it works when you copy the right kind of trader, in the right kind of market, at the right size. It fails, reliably, when you don't. This post works through the numbers so you can tell the two apart.

Copy trading copies decisions, not returns

When a wallet you follow buys a Polymarket outcome, a copy-trading bot sees the trade and places a matching order from your wallet. You get the same decision: same market, same side, similar size relative to your bankroll. You don't get the same price. Three costs sit between their result and yours:

  • Latency. You act after the trader. In the seconds between their fill and yours, their own order may have moved the price.
  • Slippage. To make sure a copy order actually fills, it's placed a little above the last price, so it crosses the spread.
  • Fees. Polymarket charges takers a fee on most market categories, and the copy-trading tool charges its own.

None of these costs is large on its own. Together they decide whether copying a trader keeps their edge or erases it.

A worked example with real fees

Polymarket's taker fee is calculated as shares × feeRate × p × (1 − p), where p is the share price. Per Polymarket's fee documentation, the rate is 0.07 for crypto markets, 0.05 for sports, economics, culture and weather, and 0.04 for finance, politics, mentions and tech. Geopolitical and world-events markets are fee-free, and makers (resting limit orders) never pay.

Take a sports market. A trader buys YES at 62¢ as a taker:

  • Their fee per share is 0.05 × 0.62 × 0.38 ≈ 1.18¢, so their all-in cost is about 63.2¢.
  • If YES wins, each share pays $1: a profit of about 36.8¢, or 58% on the money put in.

You copy the trade a moment later:

  • You fill at 63.5¢, 1.5¢ of slippage because their buy lifted the offer.
  • Polymarket's fee at that price is 0.05 × 0.635 × 0.365 ≈ 1.16¢.
  • PolyZig's fee on the Pro tier is 0.40% of 63.5¢ ≈ 0.25¢.
  • All-in cost: about 64.9¢. If YES wins, you make about 35.1¢ per share, roughly 54%.

54% versus 58% on a win doesn't sound like much. The important number is the edge, not the payout. Suppose the trader is right that YES is really a 66% outcome:

  • Their expected edge is 66 − 63.2 = 2.8¢ per share.
  • Yours is 66 − 64.9 = 1.1¢ per share. You kept about 40% of their edge.

Now suppose the trader's real edge was thinner, and YES is a 64% outcome. They are still +0.8¢ per share. You are −0.9¢: a winning trader, copied faithfully, loses you money. That's the whole story of copy trading in one line: the costs of copying are roughly fixed, so only traders whose edge per trade is bigger than those costs are worth copying.

Which traders copy well

  • Wide edge per trade. Look at profit per dollar traded (monthly P&L divided by monthly volume). A trader making 5–15¢ per dollar traded has room for your costs; one making 0.3¢ doesn't. Our best traders to copy list shows this number for every trader.
  • Long holding periods. If a trader enters days before resolution, a few seconds of latency is noise. If they flip positions within a minute, latency is everything.
  • Liquid markets. A 1¢ spread with depth behind it means your copy fills close to theirs. A thin book means their order is the price move you pay for.
  • Fee-light categories. The same edge survives better in politics (0.04 rate) or fee-free geopolitics than in crypto (0.07).
  • Consistency across time windows. Profitable this week, this month and all-time is far more convincing than one huge month.

Which traders you should not copy

  • Market makers. Huge volume, tiny profit per dollar, orders on both sides. Their edge is the spread, and as makers they pay no fees. A copier who crosses the spread and pays the taker fee is on the opposite side of that edge.
  • Arbitrage and high-frequency bots. Their profit comes from being first. A copier is, by definition, second.
  • 5- and 15-minute crypto scalpers. Crypto carries the highest fee rate, peaking at 1.75¢ per share at 50¢, which is 3.5% of the price, on every entry. On a market that resolves in minutes, that's more than most edges.
  • Whales who move the book. When a $200,000 order sweeps three price levels, the price you copy at is the one after the sweep.

Leaderboards lie by omission

Polymarket's profits are extremely concentrated. A Solidus Labs analysis reported by CoinDesk found that in politics markets from December 2025 to February 2026, fewer than 1% of wallets captured roughly half of all gains. That cuts both ways. There really are skilled traders worth following, but any single-period leaderboard also surfaces lucky wallets alongside skilled ones, and you can't tell them apart from one number.

Some behaviours also make a wallet look better, or harder to follow, than it is:

  • Merges and hedges. A trader can exit by buying the opposite outcome and merging the pair back into collateral, instead of selling. A copier watching only for sells may keep holding after the trader is flat.
  • Iceberging. A large position built from dozens of small orders. Each copy is small, and by the time you've matched the full size, the price has moved.
  • Survivorship. You only see the wallets that won this month, not the similar wallets that lost with the same strategy.

This is why our list requires a trader to be profitable over both the last 30 and the last 7 days, and ignores wallets with no reported trading volume.

Size so a mistake is survivable

Even a well-chosen trader will have losing weeks, and you will occasionally pick the wrong trader. Sizing is what keeps that from mattering:

  • Start small. Copy at 0.1x–0.25x of the trader's size, or a fixed $5–$20 per trade.
  • Cap every market. A max position per side, plus a total cap per market so a trader adding to a loser can't keep dragging you in. On PolyZig this is smart sizing.
  • Set a slippage limit. If the price has already run past your tolerance, skipping the trade is the right outcome.
  • Paper trade first. Run the copy in paper-trading mode for a week. The paper fills show you the real gap between their price and yours before any money is at risk.
  • Spread across two or three traders in different categories rather than betting everything on one.

So, does it work?

Copy trading on Polymarket works as a way to borrow someone else's research, not their execution. When a trader's edge comes from knowing something (a sports model, a read on an election, weather data), and they express it days ahead in a liquid market, you can keep a meaningful share of that edge after costs. When their edge comes from speed, rebates or size, you can't, and copying them loses money even while they win.

If you want to try it:

Prediction-market positions can go to zero, and past performance doesn't predict future results. Only copy with money you can afford to lose.