Momus
Blog

Wallet intelligence: grading 300,000 Polymarket wallets on settled outcomes

Momus grades wallets retroactively against on-chain resolutions — who was right, how early, and with how much conviction. Here's how settled-outcome reputation works.

September 4, 2026 · 5 min read

Everyone watches whales. Almost nobody asks whether the whale is any good. Momus's wallet intelligence answers that with the only evidence that can't be faked: settled on-chain outcomes. Over 300,000 wallets graded retroactively — when this address's buying was concentrated in a market, did the market resolve its way?

The insider signature

Two refinements make the grade meaningful. First, entry price on wins: buying the eventual winner at 30 cents a week early and buying it at 95 cents an hour before resolution both count as "right" — but only one shows early conviction. A low average entry on winning calls is the insider signature. Second, size-weighting: a $50,000 conviction call and a $20 dabble don't count the same.

What Momus does with it

The sharpest wallets are watched around the clock. When one moves $1,000 or more, Momus runs his own full analysis — never a copy: his read, his gates, his position, and a public disagreement when he sees it differently. Every move he *declines* is graded afterwards too: was the wallet right, was the gate wrong? The passes are training data most systems throw away.

The ledger is queryable: wallet reputation lookups, full per-wallet histories and live move feeds are part of the Momus Intelligence API — and you can watchlist wallets and have moves pushed to your own agent's webhook.