Six real-time scores that answer one question: is this wallet worth following?
Smart money. Got in early, profitable across tokens, still accumulating. When it buys, pay attention.
Exit liquidity. Late, small, loses money, already selling. A raw transaction feed makes this wallet look identical to the one on the left.
Every score: higher is better. Updated live, not daily.
Each is scored from raw on-chain swaps and transfers, then ranked against the token's own holders — recomputed at every block, impossible to fake. Here's what each one measures, with the full ladder of tiers.
Execution skill — who actually makes money
How well a wallet times entries and exits. Computed from realized PnL — average sell price vs average buy price — then percentile-ranked against every other trader in the same token and mapped to an IQ score (avg ≈ 100). Needs real exits: no sells, no score.
| Tier | IQ score | What it means |
|---|---|---|
| genius | 159–165 | Top 5% — best realized returns in the token |
| gifted++ | 146–158 | Top 15% — consistently strong exits |
| gifted | 126–145 | Top 30% — above-average timing |
| avg++ | 107–125 | Top 45% — slightly ahead of peers |
| avg | 87–106 | Middle of the pack |
| below avg | 61–86 | Bottom 60% — trailing peers |
| dumb | 42–60 | Bottom 20% — frequently exits at a loss |
| idiot | 35–41 | Bottom 5% — worst timing in the token |
Use it to follow proven winners and screen out wallets that buy high and sell low.
Per-token scores are called HolderDNA. A wallet's overall WalletDNA blends them across every token it touches, weighted toward where it's most active.
How it works
No black box. WalletDNA is computed from on-chain events with an auditable pipeline — the same scores you read are recomputed continuously as new blocks land.
Ingest & decode
Raw logs decoded into one canonical event feed — buy / sell / mint / transfer, normalized across chains.
Scheduled SQL tasks roll events up and rank wallets within each token, gated by a maturity & confidence factor.