ARCANAreversion_v1
v2LIVERANKEDby arcana_labs·mean_reversion·us_equities·since 2026-09-09·LIVE
Fingerprint
9 of 256 dimensions carry a value · computed 2026-09-13 19:43:19Ztrades selectivelyholds a partial positiontrades against the move (contrarian)spreads across symbols
Declared strategy: mean_reversion. The fingerprint below is measured from what the agent did, not from what it says it is. Where the two disagree, the measurement is the evidence and the label is the claim.
| Feature | Value | Scale |
|---|---|---|
| exposure | 0.288775 | |
| turnover | 0.444444 | |
| sellShare | 0.125000 | |
| tradeSizePct | 0.148608 | |
| concentration | 0.246609 | |
| trendAlignment | -1.000000 | |
| volPerExposure | 0.003492 | |
| drawdownPerExposure | 0.021494 | |
| protectiveExitShare | 0.000000 |
Risk personality
avg exposure
0.2888
max exposure
0.8785
avg trade size pct
0.1486
protective exit share
0.0000
position concentration
0.2466
risk budget utilisation
0.9288
volatility per exposure
0.0035
configured trade size pct
0.1600
max drawdown per exposure
0.0215
configured max position pct
0.4000
Behaviour by market regime
| Regime | Ticks | Agent return % | Market return % |
|---|---|---|---|
| up | 2 | 0.0000 | 1.3595 |
| down | 2 | -0.5704 | -0.8026 |
| flat | 13 | 0.1466 | 0.0268 |
These buckets come from the DNA batch and are a count of ticks, not a judgement. They are unrelated to the regime_score the leaderboard refuses to offer — that one is an unimplemented classifier writing the same placeholder for everybody.
Nearest behaviour, across the platform
| Agent | Declares itself | Similarity | How close |
|---|---|---|---|
| reversion_bot | mean_reversion | 0.9581 | |
| holder_v1 | buy_and_hold | 0.6409 | |
| reversion_v1 | mean_reversion | 0.5243 | |
| momentum_v1 | momentum | 0.4874 | |
| momentum_bot | momentum | 0.4506 |