Across 60 games, 48 graded wallets bought before kickoff. Below: which side each cohort favoured, how much sat behind it, and — once played — how it settled.
Hits–misses over games where a cohort had a call: at least 3 of its wallets bought the moneyline before kickoff, at least $250 between them, and at least 60% of that money on one side. Everything thinner is shown on the table but not counted.
Watching every College Football wallet by hand is the part that doesn't scale. Set one threshold and CopyGrade emails you when any wallet it tracks — on your watchlist or not — opens a buy at or above that size on College Football markets.
Each row is one College Football game and its moneyline market on Polymarket. For every game we take the buy fills that CopyGrade's graded wallets made on that market before kickoff, split them into two cohorts — clean (graded wallets that pass CopyGrade's farming check and score 65 or higher, the Candidate band) and flagged (graded wallets that carry a farming-risk flag (watch or severe), at any score) — and show which side each cohort's money favoured, how much money that was, how many wallets, and their average fill price. A clean wallet scoring under 65 is in neither cohort. Sells, spreads, totals and props are not counted. A week runs Tuesday 09:00 UTC to the next Tuesday 09:00 UTC, so a Monday-night game stays with its own week.
A cohort has a call only when at least 3 wallets, $250 and a 60% majority sit behind the lean; the ledger counts only those. A one-week record is a handful of coin-flips — read the season line, not the week, and read the guide before reading anything into it. Everything here is an observation about what a cohort bought, not a recommendation about what you should.
Wallet-level detail — who is positioned in a game and at what size — is on each game's page. Aggregates only here, by design.
As of . Leans update as pre-kickoff fills land; results update as games settle.
CopyGrade scores, verdicts, and flags are algorithmic opinions derived from public on-chain data. They may contain errors and are not statements of fact, accusations of wrongdoing, or financial advice. How scoring works · Operate this wallet? Dispute an assessment.