CopyGrade
Farming Risk

Specific behaviours, each one named.

A farmer builds a position quietly, advertises a great-looking record, attracts copiers, then uses those copiers as exit liquidity. Farming Risk is the forensic check that names the behaviour and shows the on-chain trades that triggered it. No vibes — only evidence.

Last reviewed 2026-06-01
Why farming is the threat

The gap nobody else is looking at.

Polymarket’s leaderboard ranks wallets by raw PnL. That is exactly the metric a farmer optimises for: a clean number on a public board, generated by a position that copiers will help unwind. The leaderboard rewards the behaviour it should be flagging.

The dozen copy-trading bots competing on speed cannot help. They mirror the visible trade. They cannot see an iceberg being built below their alert threshold, they cannot tell a self-trade from a real fill, and they cannot mirror a YES/NO merge — the cleanest way to exit a position without it ever showing up as a sell.

That is the gap CopyGrade fills. We name the behaviour. We show the trade.

The taxonomy

The farming taxonomy.

Each card describes the behaviour in plain English, the exact on-chain pattern that triggers a finding, and where the detector sits in v1.

Iceberg accumulation
Live in v1
A position built in small slices so no single trade looks loud enough to copy. By the time the visible trade prints, the farmer is already long.
On-chain pattern
At least 8 sub-$200 buys clustered on a single market within a 30-minute window — sized to sit under copy-bot alert thresholds.
Self-trade wash
Live in v1
The same operator on both sides of a thin market. Generates volume and the illusion of a real fill, which props up the wallet’s headline activity.
On-chain pattern
At least 8 near-price buy/sell pairs on the same market and outcome within a 10-minute window — well past the back-and-forth of ordinary active trading.
Manufactured-volume churn
Live in v1
A wallet that looks busy because it is — but the activity isn’t producing edge. Volume for the leaderboard, not for the copier.
On-chain pattern
At least 350 trades over the window producing less than 1.5% net edge relative to capital deployed.
Pump-and-dump
Live in v1
A position accumulated quietly, then exited into a price run-up the wallet’s own flow drove. The copier mirrors the entry and becomes the exit liquidity.
On-chain pattern
A ≥$500 buy cluster with the buy prices walking up, then sells at a VWAP ≥30% above the buy VWAP, the whole cycle inside 12 hours. Weighted as suggestive — it can read like skilled timing — so it nudges the level rather than hard-vetoing.
Copy-bait
Live in v1
The inverse of an iceberg: build quiet, then advertise loud. A conspicuous, mirror-me entry sized to be copied, then a quick exit once followers pile in.
On-chain pattern
At least 4 quiet sub-$200 buys, then a single ≥$2,000 headline entry, then an exit within 6 hours on the same market.
YES/NO stealth merge
Live in v1
Offsetting YES and NO positions on one market — a market-neutral complete set (1 YES + 1 NO redeems for $1). The visible directional buys get mirrored as conviction, but the wallet is hedged and can merge the set back to cash, risk-free, without it ever showing up as a sell — leaving the copier holding the one-sided exposure.
On-chain pattern
A wallet simultaneously holding ≥$500 of offsetting YES and NO shares on the same binary market, the two legs at least half-balanced. Weighted as suggestive (it can be legitimate arbitrage or market-making), so it nudges the level rather than hard-vetoing.
Decoy wallet clusters
Live in v1
A primary wallet and a handful of linked sub-wallets trading opposite sides. The cluster has near-zero net edge, but the primary shows a clean record because the losses live on the decoys.
On-chain pattern
A primary plus up to five sub-wallets that share ≥3 markets, co-trade within 5 minutes, take opposite sides, and net near-zero aggregate edge. Detected by a cross-wallet pass and folded into the primary's farming level. Wallets sharing an on-chain funder — their earliest USDC.e depositor on Polygon — are linked as one operator, the strongest signal, corroborating the timing + opposite-side correlation.
Cross-wallet coordination
Live in v1
Several wallets trading the same markets in tight time-lockstep without the decoy structure — pump rings and wash networks that move price together, then distribute the exit to followers.
On-chain pattern
Three or more wallets co-trading several of the same markets within 5 minutes of each other. Weighted as suggestive (correlation, not proof), so it nudges the level rather than hard-vetoing.
How findings flow into the score

The farming veto.

Detector count maps directly to farming level, and farming level caps the Copy Score. A wallet that farms its copiers can never read as a strong candidate.

Signals
Clean
No cap
No detector tripped. The Copy Score reflects the full five-factor model.
Signals
Watch
Capped at 64
A moderate-confidence finding (or a couple of borderline ones). The wallet can never read as a strong copy candidate until it clears.
Signals
••
Severe
Capped at 30
A high-confidence signature, or several moderate ones by weighted severity. The score is vetoed to the Avoid band, however good the headline PnL.
On the verdict page

Every flag links to the trade.

When a detector triggers, the Copy Verdict shows the signature by name, the market it fired on, and the recency. The detail line cites the exact trade count, window, and threshold that produced the finding. Every flag shows the concrete on-chain evidence behind it — the trades themselves, not a black-box score.

If you disagree with the finding you can audit it. Every signal points back to the on-chain trades it was built from. No black-box scoring, no hand-waving, no “our model thinks.”

Limitations

What this version doesn’t do yet.

The thresholds — 8 buys, 30 minutes, 8 pairs, 350 trades, 1.5% edge — are heuristic priors, tightened in a first pass against real Polymarket activity (the synthetic-tuned originals over-flagged ordinary active traders). We snapshot every score and measure forward copier outcomes against it to recalibrate them — a propose-only loop that is still warming up, with the first retune waiting on a 30-day outcome window.

False positives are possible. A genuine scalper running tight spreads can clear the wash threshold without intent. A patient accumulator can look like an iceberg. The verdict cites the trades so the reader can decide.

Decoy wallet clusters are flagged by a cross-wallet pass — timing + opposite-side correlation with near-zero aggregate edge across the cluster, now corroborated by shared-funder linkage (wallets with a common on-chain funder are likely one operator). The funder signal uses a low-fan-out filter, so wallets funded through shared infrastructure — exchanges, bridges, relayers — aren't linked; a cluster funded from separate sources that only coordinates loosely can still slip through.

YES/NO stealth merges are now flagged: a wallet simultaneously holding offsetting YES and NO positions on one market is holding a market-neutral complete set it can merge back to cash without a visible sell. Because the same shape can be legitimate (sub-$1 YES+NO arbitrage, market-making), it's weighted as suggestive — a lone merge nudges to watch rather than hard-vetoing, and the verdict cites the offsetting positions so the reader can judge.

See it on a real wallet.

The mock dataset contains a known farmer. The verdict page shows the signals firing, the trades they were built from, and the score the veto produced.