How to Copy Trade on Polymarket: Mechanics, Sizing, and Day-One Settings

Last updated 2026-07-15 · methodology · changelog

Copy trading on Polymarket means a bot watches a chosen wallet on-chain and mirrors its trades into yours. The mechanics are simple; the details — sizing mode, latency, and the limits you set on day one — decide whether the copy resembles the original.

What copy trading actually is

Every Polymarket position lives on Polygon, so any wallet’s entries and exits are public the moment they hit the chain. A copy-trading bot subscribes to that feed: you pick a wallet address, the bot detects each new trade it makes, and it places a corresponding order from your own balance. The leader never knows and never has to consent — the blockchain is the API.

Execution differs by tool. PolyBot documents up to three active copy subscriptions per wallet, with sells copied proportionally — if the leader exits 40% of a position, your position trims 40%, not the same dollar amount. Kreo advertises instant mirroring of top Polymarket and Kalshi wallets with alerts within seconds. PolyCop reports that roughly 30% of its copies fill in the same Polygon block as the source trade. None of these latency figures are independently audited, which is worth remembering whenever speed is the selling point.

Sizing modes: fixed, proportional, Kelly

The sizing mode is the single most consequential setting, because it decides how the leader’s conviction translates into your risk.

Fixed sizing puts the same amount on every copied trade regardless of what the leader wagered. It’s the easiest to reason about and the best fit for small bankrolls — PolyBot’s fixed mode starts at $1.10 per copied trade — but it flattens the leader’s signal: their high-conviction bets and throwaway punts get identical weight in your account.

Proportional sizing scales each copy to the leader’s stake — PolyBot documents multipliers from 0.01x to 10x, and Kreo offers percent-of-balance sizing. This preserves the leader’s conviction structure, but a whale’s “small” $5,000 position at 0.01x is still $50, so check the arithmetic against your balance before enabling it.

Kelly-fraction sizing, documented by PolyCopy, sizes each copy from an estimate of the trade’s edge relative to your bankroll. It’s the most theoretically grounded mode and the least forgiving of bad inputs: if the estimated edge is stale, Kelly confidently sizes into a mistake.

Latency: why your fill is worse than theirs

Between the leader’s trade confirming on-chain and your copy reaching the order book, the price moves — often because of the leader’s own order, and sometimes because dozens of other copiers are hitting the same book. On liquid markets the gap may be a fraction of a cent; on thin ones it can swallow the edge entirely.

The documented spread across tools is wide. PolyCop advertises same-block or next-block fills. At the other end, independent research on PolyGun recorded a user complaint of copies landing 30 to 60 seconds behind the source activity — long enough for the market to move against them. This is why serious copy engines expose a slippage tolerance or price offset: a rule that skips the copy when the price has already drifted too far from the leader’s entry. A skipped trade costs you nothing; a chased one can cost more than the fee.

Risk controls to set on day one

Before the first copy executes, set the limits — not after the first surprise. The deepest documented control set we verified is PolyBot’s (our methodology explains how we compare them), but the core four exist in some form across the top tools:

• A daily spending cap. An active leader can fire dozens of trades a day; a cap converts “whatever they do” into a bounded number.
• Per-trade maximums. So one outlier position can’t dominate your book.
• Odds-range filters. PolyBot documents restricting copies to, say, 20–80¢ outcomes — skipping longshot lottery tickets and near-resolved markets where fees and slippage dominate.
• Category filters. Kreo documents 13 market-category filters and PolyBot documents white/blacklists with Sports and Crypto subcategories, so you can copy a trader’s politics book without inheriting their 3 a.m. crypto habit.

Then attach exits. Take-profit and stop-loss on copied positions are documented by PolyBot, Kreo, and PolyCop; PolyBot and Kreo add trailing stops. Treat every stop as a trigger plus an execution attempt rather than a guaranteed floor — PolyBot’s own docs say exactly that, and the candor applies to the whole category.

Choosing the tool and the target

Two decisions remain: which bot, and which wallet. For the first, our ranked comparison of Polymarket copy-trading bots weighs custody, fees, filters, and track record. For the second — the harder problem of separating skill from survivorship on leaderboards — see the companion guide on choosing wallets to copy. Start with one leader, small fixed sizing, tight caps, and a week of watching the copy log before scaling anything.