Signal Lab
Alert outcomes, precision and backtests
Tracks how the engine's alerts have performed, surfaces precision/recall summaries across windows, finds tokens that look like a seed token and runs simple wallet copy backtests. The model gets sharper as the outcome ledger grows.
Mock signal lab · mock · refreshed 45m ago · TTL 300s
Precision summary
How well are alerts performing?
Precision = hits / scored. Recall is approximated from the hit-vs-miss balance because true negatives aren't observable for alerts.
Last 1h
0%
precision · recall proxy 0%
- Hits
0
- Misses
0
- Noisy
0
- Pending
1
Med hit return
0%
Med miss return
0%
Last 24h
36%
precision · recall proxy 50%
- Hits
5
- Misses
5
- Noisy
4
- Pending
2
Med hit return
+27%
Med miss return
-18%
Last 7d
36%
precision · recall proxy 50%
- Hits
5
- Misses
5
- Noisy
4
- Pending
2
Med hit return
+27%
Med miss return
-18%
Alert outcomes
Recent fired alerts and what happened next
Outcome band summarises whether the alert led to a directional move, a noise event, or a miss. Pending entries have not been scored yet.
| Outcome | Subject | Rule | Realised | Alert conf. | Outcome conf. | Fired |
|---|---|---|---|---|---|---|
| pending | WIF | Smart-money inflow > 100K (1h) | pending | 68% | — | 3m ago |
| pending | WIF | Holder concentration > 50% | pending | 67% | — | 1h ago |
| hit | PEPE | Smart-money inflow > 100K (1h) | +15.0% | 88% | 88% | 2h ago |
| noisy | PEPE | Holder concentration > 50% | +1.0% | 87% | 87% | 3h ago |
| hit | BONK | Smart-money inflow > 100K (1h) | +27.0% | 88% | 83% | 4h ago |
| noisy | BONK | Holder concentration > 50% | +1.0% | 87% | 82% | 5h ago |
| miss | AGENT | Smart-money inflow > 100K (1h) | -24.0% | 83% | 63% | 6h ago |
| miss | AGENT | Holder concentration > 50% | -23.0% | 82% | 62% | 7h ago |
| miss | GRIFT | Smart-money inflow > 100K (1h) | -11.0% | 76% | 66% | 8h ago |
| miss | GRIFT | Holder concentration > 50% | -10.0% | 75% | 65% | 9h ago |
| hit | MEME2 | Smart-money inflow > 100K (1h) | +17.0% | 88% | 93% | 10h ago |
| noisy | MEME2 | Holder concentration > 50% | -1.0% | 89% | 94% | 11h ago |
| noisy | RUGZ | Smart-money inflow > 100K (1h) | +2.0% | 68% | 68% | 12h ago |
| miss | RUGZ | Holder concentration > 50% | -18.0% | 67% | 67% | 13h ago |
| hit | ARBI | Smart-money inflow > 100K (1h) | +43.0% | 90% | 75% | 14h ago |
| hit | ARBI | Holder concentration > 50% | +42.0% | 89% | 74% | 15h ago |
Pattern similarity
Tokens that look like WIF
Similarity is matched against historical mocks. Outcome is shown so you can see how comparable patterns played out.
- PEPE90% similarsuccessful
Shares meme characteristics with WIF.
- BONK79% similarsuccessful
Shares meme, solana characteristics with WIF.
- AGENT68% similaralive
Comparable on chain (base) and recent flow profile.
- GRIFT57% similarrugged
Comparable on chain (base) and recent flow profile.
- MEME246% similaralive
Shares meme characteristics with WIF.
Wallet backtest
Hypothetical copy outcomes — Trailing 30 days
Per-wallet simulated returns assuming the listed copy delay. Drawdown is the worst peak-to-trough move during the window.
| Wallet | Copy delay | Realised | Worst drawdown | Trades |
|---|---|---|---|---|
| ICEBERG | 60s | +29.0% | -19.0% | 21 |
| ALPHASEER | 90s | +9.5% | -15.0% | 27 |
| OXSMART | 120s | +3.0% | -10.0% | 16 |
| BASEBEACON | 150s | +6.5% | -15.0% | 29 |
| SILENTGAINS | 180s | +21.0% | -19.0% | 17 |
| ARBIOWL | 210s | -3.5% | -8.0% | 28 |
| ETHERELF | 240s | -6.0% | -23.0% | 23 |
| FRESHHAND | 270s | +15.0% | -15.0% | 15 |
- · Backtest assumes ideal execution (no slippage past the modelled rate).
- · Copy delays are fixed per wallet — real-world delays vary widely.
- · Past behaviour does not predict future behaviour.
Research output only. Backtests assume ideal execution and do not predict future outcomes.
Research and analytics only. Smart Money never holds your funds, signs transactions, or executes trades. Users always sign in their own wallet.