Copy trading and the entry-gap problem

Polymarket positions are public. You can watch a wallet that has made a lot of money, see what it just bought, and buy the same thing. This is the most intuitively appealing strategy in the market, and it fails for a reason that has nothing to do with whether the trader is good.

You do not get their price

They bought at 66c. By the time the fill is on-chain, indexed, and in front of you, the market is 74c — partly because their own buying moved it. You are not taking their bet. You are taking a different, worse one:

Same trade, same trader, same opinion — profitable for them, unprofitable for you. Nothing about their skill has changed. Eight cents of entry gap did all of it.

Survivorship: the leaderboard is a filter

A profit leaderboard shows the wallets that survived. It cannot show the ones that took identical risks and busted, because they are not on it any more. With enough participants, some are up a great deal on variance alone, and their behaviour is indistinguishable from skill in the only data you have.

Before copying anyone, ask: how many bets is that record over? Is the profit spread across many positions or one lucky call? Are they trading a category where public information is enough, or do they plausibly know something? A wallet up 400% on six trades is a coin that landed heads six times.

When following is actually reasonable

• The gap is small. If they paid 66c and it is 67c, you are taking essentially their bet. Check the gap first, every time.
• Long-dated markets. A position resolving in six months has time to be right regardless of an entry point a couple of cents worse.
• Long records. Hundreds of positions across categories beats a spectacular short run, every time.
• As a shortlist, not a signal. Their trades tell you which markets are worth your own look. What you do after that look is your decision, at your price.

Test it before it costs anything

Copying is an empirical question with a cheap answer: mirror the wallet into a paper account at the prices you would actually have got, and look at the record after a few dozen positions. If it works, it will still work with real money. If the entry gap eats it, you will have found that out for nothing.