Polymarket Copy Trading: Following 30 Seconds Late Cost 1.43¢ a Share, Before Fees

Copy trading on Polymarket copies a trader's decisions, not their returns, and the price moves before the copy arrives. Vultax followed 54,828 buys by the 1,000 most profitable wallets in its Polymarket cohort on 12 and 13 September 2026. Across 17,088 of them in event markets, a follower 30 seconds late paid 1.43 cents a share more on average, and paid more than the leader 57.9% of the time. Ten minutes late, the extra was 1.29 cents: most of the cost was already in the price within the first half-minute.

Fees and fill add to it. Worked from the published fee formula, a copier who follows a 100-contract buy at 50 cents in a crypto market is down about $3.00 on a $50 stake, 6%, before the market moves at all, counting the taker fee, a 0.5% platform fee and a fill one cent worse. Politics markets cost about 4.5%; geopolitics, which Polymarket does not charge fees on, about 2.5%.

The tools differ mainly in custody and automation. Polycopy is non-custodial, free to follow and paper-trade, $30 a month to run live bots, and charges 0.5% per executed trade. LuckyLobster runs agents on managed wallets. Laika Labs and Alphascope are leaderboard-and-alert products that leave execution to you.

Four products that let you follow a Polymarket wallet, plus the simulator, read from each vendor's own pages on 5 September 2026.

Download table figures:CSVJSON with sources

Polymarket's taker fee is C × feeRate × p × (1 − p), where C is contracts and p the price; makers pay nothing. The rate is 0.07 on crypto, 0.05 on sports, economics, culture and weather, 0.04 on finance, politics, mentions and tech, and zero on geopolitical markets. At p = 0.50 the fee is at its maximum, so the worked example below is the worst case at the most common price. The Polymarket fee calculator applies it to any size, price and category.

Suppose the leader buys 100 Yes contracts at 50 cents in a crypto market. Their stake is $50. Your copy, placed as a market order a few seconds later, pays three things the leader did not.

Download table figures:CSVJSON with sources

Download chart figures:CSVJSON with sourcesSVG chart

• The leader paid none of these if they posted a limit order - "Makers are never charged fees", so a leader who rests orders keeps the 3.5% you paid to take.
• The costs repeat on exit - the same three land again when the position is closed or auto-closed, so a round trip in a crypto market costs a copier roughly 12% of stake at 50 cents.
• Speed helps only if it is measured - a sub-50 ms mirror still takes liquidity and still pays the fee; it reduces drift, not the other two costs, and in the measurement below most of the drift was already in the price 30 seconds after the leader's trade.

The worked example assumes a fill one cent worse than the leader's; Vultax measured the move instead. The leaders are the 1,000 wallets with the highest observed profit in its cohort of 80,000 Polymarket traders as refreshed on 12 September; the 1,000th had an observed profit of about $154,000. Of the 1.86 million trades recorded between 15:06 UTC on 12 September and 22:55 UTC on 13 September, 54,828 were the leaders' buys at prices between 3 and 97 cents, by 195 of the 1,000 wallets.

The follower's price is the first buy of the same outcome by any other wallet at or after the delay, provided one happened within five minutes of that mark; the leader's own later buys were skipped, so a leader splitting an order does not count as its own follower. Up-or-down markets were measured at 30 seconds only, because a five-minute market has usually settled before a ten-minute delay ends.

In event markets, everything except the up-or-down crypto series, the follower paid 1.43 cents more per share at 30 seconds and at two minutes, and 1.29 cents at ten, with a median extra of about 1.4% to 1.7% of the leader's price at every delay. That pattern fits a price that moves at or just after the leader's own trade, before any realistic copier arrives. In up-or-down markets the delay hardly mattered: 0.42 cents, with the follower paying more 52.9% of the time, close to a coin flip.

A trade recorded at that later price does not show that your own quantity would have filled there, and a buy with no other buyer in the five minutes after the mark is outside the estimate. A copier's own record needs the missed trades beside the filled ones: a table of successful entries can describe a different strategy from the one an alert would have let you follow.

Download table figures:CSVJSON with sources

Split by rank, the gap widens at the top. Behind the 100 most profitable wallets, a follower paid 1.97 cents more after 30 seconds and 2.39 cents after ten minutes, and paid more on 60.6% of 30-second follows. Behind wallets ranked 101 to 1,000 the figures were 1.20 and 0.80 cents.

One reading is that the top wallets trade earlier on information that the market then prices in; another is that their orders are larger and move thin books further. This study cannot separate the two, but both mean the same thing for a copier: the traders worth copying are the ones whose trades leave the least on the table.

Download chart figures:CSVJSON with sources

The case for copying rests on persistence, and most winners do not persist: across all wallets active in consecutive months from March 2025 to March 2026 (431 million fills, 13 month pairs), a profitable month was followed by another 41–51% of the time. The active few are the exception. Among the 184 wallets with ten or more trades in both June and July 2026, 124 of the 137 June winners were profitable again in July, and the rank correlation between months was 0.83. But two facts from the same data cut against the copier.

First, one-trade wallets were profitable 12.2% of the time. The wallets worth copying are the busy ones, and busy leaders multiply the costs above by their trade count. Second, gains are concentrated: the top 1% of winning wallets, 36 addresses, took 43.9% of all realised gains over 18 June to 14 July 2026, and the top 10% took 79.5%. Inside that decile the leaders are the always-on wallets that fill hundreds of trades a day, each of which would charge you the table above.

Rules that follow from the numbers.

• Paper-trade the leader for a month first - Polycopy's free bots do this; if the simulated copy is not clearly positive net of the fee formula, the live copy will not be either.
• Copy with limit orders, not market orders - resting at the leader's price or better converts the taker fee to zero and removes the worse-fill cost; you will miss some entries, and the entries you miss are the ones that moved against you fastest.
• Size the copy to the book - in the order-size and depth audit, 100 contracts filled at or near the quote almost everywhere, while at 10,000 more than a third of books could not fill the order from what they displayed.
• Prefer fee-free categories - geopolitical markets carry no Polymarket fee, so the copier's structural disadvantage there is the platform fee and the fill alone.
• Follow decisions, not wallets - the useful output of a leader is the market they chose and the side; entering later, with your own size, at a resting price, keeps the information and drops most of the cost.
• Never copy a one-trade wallet - the persistence evidence does not reach it.

The strip at the top of this page is rewritten every ten minutes from Polymarket's public leaderboard by day, week and month, its data API's trades of $10,000 or more, its global open interest, and Vultax's record of every fill it observes, including the always-on wallet cohort. The Vi IQ beneath it scores the six domains the Vultax terminal scores for a pair, computed for this page's subject from published inputs, with a domain that cannot be computed shown as unavailable rather than filled in. The figures in the body of the article are the measurement it was written on and are dated; the strip is what the same measurement says now.

The strip shows the median trade Vultax observed in the last day, the share of wallets that traded exactly once, the number of trades of $10,000 or more in the last hour and how many wallets traded in the last day. The first two describe the crowd a copier joins; the third is the flow a copier is late to; the fourth is the size of the pool.

The three costs have a fourth cousin: the leader may not be who the copier thinks. Columbia researchers flagged about a quarter of Polymarket's volume as wash trading and roughly one wallet in seven, many with no profit, which is what a wallet farming a ranking looks like from the outside; and the CFTC's February 2026 advisory cites insider trading by a candidate on his own race and by a video editor on his channel's contracts. A copier of the first pays to follow noise; a copier of the second follows conduct the Commission prosecutes.

The venues are also changing what can be copied. Polymarket's perpetual futures, launched 3 September with up to 20x leverage and closed to U.S. traders, mean a leader's profit may soon include leveraged positions that a copier on the U.S. exchange cannot hold; and with Kalshi at about 90% of combined contract volume in late August, the leaders on the two venues are increasingly different people.

Copy-trading vendors publish their own research, for example Polycopy's report on copy trading. When comparing claims, check whether they measure the delay against a real follower's next fill, as the Vultax measurement above does, or against the leader's own price.

Tool facts were read from each vendor's own pages on 5 September 2026; where a vendor site blocked automated reading the cell says so and leans on third-party guides, which is weaker evidence. The fee arithmetic uses Polymarket's published formula and rates as of that date and a 0.5% platform fee taken from Polycopy's pricing page; other platforms' fees are not stated. The persistence and concentration figures come from Vultax's study of Polymarket traders and carry its short windows.

Delay trades come from Vultax's record of Polymarket's public trade data, copied every two hours from a rolling buffer that does not guarantee every trade. Observed profit is each wallet's all-time figure in the cohort, not profit over the study window. Up-or-down markets are those titled "Up or Down" in Vultax's catalogue. A price one cent higher on a one-cent tick counts as a full cent even when the true midpoint moved less, which is why the table uses averages. The window, 31.8 hours over a weekend, is short.

The delay figures cover the delay alone, on trades where a follower could be observed, and understate the full cost of copying. No wallet is identified, and nothing here is a recommendation to copy any wallet or to trade. Challenge a cell at [email protected]; material changes are logged in the changelog.

• Quiet markets - at ten minutes, 39,101 of the leaders' buys had no other buy of the same outcome within the following five minutes and were left out of the delay figures, so that sample leans toward busier markets.
• Fees - the delay figures exclude the taker fee a copier pays and the leader may have avoided; the worked table above adds it.
• Size - the follower price is what one other buyer paid, not what a larger copy order would pay after walking the book.
• Exits - only buys were followed. Getting out late has its own delay cost.