Polymarket copy-trading glossary
Every term you'll meet vetting a Polymarket wallet or pointing a copy-bot at one — defined in two or three sentences each, in plain English. Definitions are grouped by topic; each one stands alone, and links go to the deep-dive that operationalises the term.
Market mechanics
Prediction market. A market where shares pay $1 if an outcome happens and $0 if it doesn't, so a share's price reads as the market's implied probability — a YES trading at 62¢ implies roughly a 62% chance. Polymarket is the largest crypto-settled prediction market.
CLOB (central limit order book). The order-matching system Polymarket runs: limit orders rest in a book, matching happens off-chain via Polymarket's operator, and settlement happens on-chain on Polygon. The book's depth at any price decides how much size can fill there — which is why a copier's fill is worse than the leader's. See how Polymarket works.
Outcome shares (conditional tokens). The ERC-1155 tokens that represent YES/NO positions, minted through the Conditional Token Framework (CTF). One YES plus one NO of the same market is always redeemable for $1 of USDC.
Split / merge. The two CTF primitives: splitting turns $1 of USDC into a YES+NO pair; merging turns a YES+NO pair back into $1 — with no trade on the order book. A merge is how a trader exits a position without ever printing a sale, which makes it both a legitimate tool and a copy-bot blind spot.
Resolution (UMA optimistic oracle). How a market's outcome becomes official: anyone proposes a resolution by posting a bond; unchallenged proposals stand after a dispute window, disputes escalate to a token-holder vote. Until resolution, capital stays locked in the position.
Maker / taker. A maker rests limit orders in the book; a taker crosses the spread to fill immediately. Copy-bots are structurally takers — they chase fills after the leader moves — so they pay the spread the leader may have earned.
Copying & costs
Copy trading. Automatically mirroring another wallet's trades, scaled to your capital. Polymarket has no native copy feature; third-party bots watch a target wallet and replay its trades for a fee. The full loop is in the complete guide.
Headline return. A wallet's advertised performance, computed on its own fills at its own prices with no copying costs. It is the ceiling a copier never reaches — typically by 20–40%.
Realistic (post-fee) edge. What a copier would actually keep per unit staked after fees, slippage, and latency — the number the Copy Score prices instead of the headline. In our June 2026 snapshot the median active wallet's realistic edge was −0.8%.
Slippage. The gap between the price you saw and the price your order actually filled at, caused by your size walking the order book. It always runs against the taker.
Latency (copy delay). The time between the leader's fill and yours — detection, bot reaction, and order placement. The price drifts toward the leader's direction in that gap, so edge decays with every second of delay.
Drawdown. The peak-to-trough loss a wallet has lived through, and the single best number to size a copy from: a 50% loss needs a 100% gain to recover. The arithmetic is in position sizing.
Sharpe-style (risk-adjusted) return. Return divided by the volatility of returns — the measure that separates a strategy from a lucky streak. Two wallets with identical PnL are not equally copyable if one bet the farm to get there.
Farming & forensics
Farming (copy farming). Building an attractive-looking trading record to harvest followers — then exiting into their buying. The umbrella term for every behaviour below; CopyGrade treats a strong finding as a score-capping veto.
Wash trading. Trading with yourself (directly or between coordinated wallets) to manufacture volume, win rate, or PnL. An independent Columbia working paper estimates roughly 25% of Polymarket's all-time volume is consistent with it.
Iceberg accumulation. Building a position in many small slices — classically under $200 each — to stay below copy-bot and whale-alert thresholds until the position is complete and the record can be advertised.
Decoy wallet cluster. A flagship wallet kept clean while linked sub-wallets (same funding source, correlated timing) absorb losses or take the other side. Visible in funding-graph analysis, invisible in the flagship's own history.
Stealth merge. Exiting by buying the opposite outcome and merging the pair to USDC instead of selling — no sale appears on the book, so bots mirroring visible trades keep holding what the leader already left.
Copy-bait. The end-to-end farming pattern: accumulate quietly, advertise a clean record, attract copiers, exit into their demand. The most dangerous wallets are skilled traders running it — skill makes the bait convincing.
Exit liquidity. What a farmed copier becomes: the buying pressure a farmer sells into. If you can't tell who the exit liquidity is, it's you.
Airdrop farming. Manufacturing activity to qualify for an expected token distribution — a major driver of Polymarket wash volume since the POLY token was confirmed in October 2025.
Scoring & vetting
Copy Score. CopyGrade's 0–100 grade of whether a wallet is worth copying, built from five factors — edge authenticity, risk-adjusted return, drawdown resilience, consistency, and the farming veto. How it's built; a score of 75+ marks the strong-candidate band, which only about 1.4% of active wallets reach.
Farming veto. The rule that a strong forensic finding caps the Copy Score outright, no matter how good the other factors look — because a convincing record is exactly what a farmer builds.
Verdict band. The plain-English compression of the score: strong copy candidate, copy with caution, or avoid. Bands are public; the numeric score and the evidence underneath are part of the full verdict.
Activity gate. The minimum recent history (20 trades in 90 days) below which a wallet isn't judged at all — every statistic computed on a thinner record is noise. The first step of the vetting checklist.
CopyGrade is analysis-only — it never executes trades or holds funds. Not financial advice.