We scored the Polymarket leaderboard. 72% are farming-flagged.
Polymarket's leaderboards are the obvious place to find a wallet to copy — they rank the biggest winners by raw profit and volume, and a copy-bot will happily point at the top of the list. So we scored it. We took the 193 wallets currently sitting on Polymarket's profit and volume leaderboards and ran each through the same Copy Score we apply to everyone. Just three of them — 1.6% — cleared the bar for a wallet genuinely worth copying. Seventy-two percent carry a farming flag. The leaderboard isn't a shortlist of good traders to copy; it's the single dirtiest slice of the platform we've measured.
A note on the sample: 193 distinct wallets that appear on Polymarket's public profit/volume boards (7d, 30d, and all-time windows) and fall under CopyGrade coverage, scored from their public trade history. The wider "active set" figures below cover 1,297 scored wallets — roughly triple our first edition's 385, as coverage has grown. All numbers are a snapshot from June 2026 and will move as coverage widens and the model is recalibrated in public.
The top of the leaderboard is dirtier than the field
The intuition is that the leaderboard skims off the best — that whatever's wrong with the average wallet, the names at the top must be cleaner. The data says the opposite. Scored side by side, the leaderboard cohort is worse than the full active set on almost every test that matters:
| Full active set (1,297) | Leaderboard top (193) | |
|---|---|---|
| Median Copy Score | 35 | 30 |
| Farming-flagged | 53% | 72% |
| — high-confidence (severe) | 43% | 57% |
| Negative realistic edge | 70% | 60% |
| Worth copying (passes all four tests) | 2.2% | 1.6% |
More farming, a lower median score, and fewer copyable wallets than the population it was skimmed from. Ranking by raw profit doesn't filter for the wallets you'd want to copy — if anything it concentrates the ones you'd want to avoid, because the behaviours that top a profit board (huge size, relentless volume, a quietly-built clean-looking record) are the same behaviours a farmer uses to attract copiers.
It's not that they can't trade — it's that they're farming you
Here's the part that's easy to get wrong, so we'll be precise about it. The leaderboard names are, if anything, better raw traders than the field: only 60% have a negative realistic post-fee edge, versus 70% across the full set. More of them can actually pick winners. That's exactly what you'd expect from the people topping a profit board.
What disqualifies them is farming, not skill. 72% carry a farming flag and 57% a severe, high-confidence one — iceberg accumulation, self-trade wash, decoy clusters, stealth merges, pump-and-dump, copy-bait. A wallet that can genuinely trade and is running a copy-bait pattern is the most dangerous profile on the platform: it has the real edge to top the board and the intent to turn the followers that edge attracts into exit liquidity. Skill makes the bait convincing; it doesn't make the wallet safe to copy. The median leaderboard wallet still ends up at a −3.0% realistic edge once you account for what a copier actually keeps — and only 1.6% survive every check.
The field behind it: 1,297 → 28
Zoom out to the full active set and the base rate is just as brutal, which is why the leaderboard's failure isn't an anomaly. Of 1,297 scored wallets, the median Copy Score is 35 out of 100, two-thirds sit below 50, and only 28 (2.2%) reach the strong-candidate band. Strip the headline number back to a fee-adjusted edge and the median wallet is slightly negative (−0.8%) before a copier adds a cent of slippage — only a quarter are positive at all.
Here's how the full set fares against each test a copier should actually apply. Each row is how many pass that single criterion; the last is how many pass all four at once.
| Test | Wallets passing | Share |
|---|---|---|
| Scored | 1,297 | 100% |
| Enough recent history to judge (≥20 trades / 90d) | 995 | 77% |
| Farming-clean | 606 | 47% |
| Realistic post-fee edge above 1% | 214 | 16% |
| Copy Score ≥ 75 | 28 | 2.2% |
| All four simultaneously — worth copying | 28 | 2.2% |
A detail worth pausing on: the 28 wallets that score ≥75 are exactly the 28 that pass all four tests. A strong Copy Score already encodes clean-history, active, and genuinely-profitable — you can't reach the top band while farming or while bleeding edge. Loosen the bar to merely clean, active, and profitable (drop the score requirement) and you still only reach 48 wallets, 3.7%. Either way, the actively-traded Polymarket population is overwhelmingly not copyable.
One honest note on the trend: our farming rate rose from 37% in the first edition to 53% here. That's not the model getting harsher — it's composition. Tripling coverage pulled in the leaderboard cohort, and that cohort is 72% flagged, so it drags the whole-set rate up. The model and its thresholds are unchanged; the population got dirtier as we sampled closer to the top of the board.
Why a profit rank can't see any of this
Raw profit is the one statistic that tells you almost nothing about whether copying a wallet makes you money. It rewards account size, reckless sizing, and luck, and it's computed on the leader's own fills — their prices, their timing, zero fees. A copier arrives a step behind, pays taker fees in and out, and eats slippage and latency. And farming is invisible to it entirely: a self-trade wash or a stealth merge or a copy-bait setup is a behavioural pattern in the trade history, not a number in the PnL column. The leaderboard sorts on the one field that hides the two things — post-fee edge and farming intent — that decide whether a copy is profitable.
What this means if you copy
The leaderboard is a list of candidates, not a list of answers — and at the very top, the base rate is worse, not better. Before you point a bot at any wallet, three checks separate the 1.6% from the rest:
- Post-fee edge, not raw profit. Ask what a copier would actually keep after fees, slippage, and latency — the headline number is a ceiling you'll never reach.
- Farming forensics. A clean-looking record is exactly what a farmer builds, and a skilled farmer builds the most convincing one. Check the trade history for the behavioural signatures, not just the PnL.
- Risk-adjusted return. A streak and a strategy produce the same headline; only one survives the next regime.
That's the work the Copy Score does on every wallet. Our graded Polymarket leaderboard puts each board name next to its Copy Score so you can see the rank-versus-reality gap directly, and the per-wallet Copy Verdict shows the post-fee edge, the farming check, and the risk-adjusted breakdown in one place — so you can tell which side of the 1.6% a wallet sits on before you commit capital.
How we got these numbers
The figures above are a point-in-time snapshot (June 2026): 193 wallets on Polymarket's public profit/volume leaderboards and 1,297 actively-traded wallets under CopyGrade coverage, each scored from its public Polymarket trade history by the current Copy Score model and recomputed on every sync. Edge figures are net of modelled trading costs and clamped to a sane range; we quote medians, which the long negative tail doesn't distort. The model is tuned against real outcomes and changes in public, with a dated note, so these numbers will shift as it's recalibrated and as coverage grows. CopyGrade is independent and analysis-only — it never executes trades or holds funds, and these scores are our documented opinion, not a statement of fact about any trader. Not financial advice.