metrics
updated 3m ago. every call graded against on-chain truth.
Total scans
12.6K
12.571 lifetime
Unique tokens
11.1K
11.086 lifetime
RUG accuracy · +24h
91.3%
resolved (94%) · ±1pp
SAFU accuracy · +24h
92.3%
resolved (93%) · ±2pp
Headline
Verdict accuracy · +24h
How often each verdict's call held at the selected checkpoint.
| Verdict | Accuracy |
|---|---|
| SAFU | 92% |
| RISKY | 35% |
| RUG | 91% |
Composition
Held vs failed · +24h
What share of each verdict's calls held up at the selected checkpoint.
| Verdict | Held vs failed |
|---|---|
| SAFU | 92% |
| RISKY | 35% |
| RUG | 91% |
Score health
Score calibration · +24h
Rug rate by score bucket at the selected checkpoint.
| Score | Rug rate · +24h |
|---|---|
| 0-9 | 1% |
| 10-19 | 5% |
| 20-29 | 11% |
| 30-39 | 27% |
| 40-49 | 31% |
| 50-59 | 36% |
| 60-69 | 39% |
| 70-79 | 39% |
| 80-89 | 42% |
| 90-99 | 37% |
| 100+ | 80% |
Reference
Decision matrix
Priority cascade for declining a checkpoint. First match wins, top to bottom.
| # | Pattern | Reason tag | Result |
|---|---|---|---|
| 1 | Token rugged, dumped, or abandoned | structural | DECLINED |
| 2 | Price down ≥ 50% | severe_price | DECLINED |
| 3 | Liquidity down ≥ 50% | severe_liq | DECLINED |
| 4 | Holders down ≥ 25% | severe_holders | DECLINED |
| 5 | Volume down ≥ 90% and under $1k | severe_volume | DECLINED |
| 6 | Any 2 of: price/liq/holders down ≥ 25%, volume down ≥ 50% | multi_signal | DECLINED |
| 7 | Otherwise | held |
By window
Miss rate over time
Of tokens that ended in downside, the fraction we called SAFU/RISKY.
| Window | Miss rate |
|---|---|
| +6h | 15% |
| +24h | 16% |
| +7d | 16% |
| +30d | 0% |