# Paper-Live Regime Comparison

Timestamp: 2026-05-14T00:20Z

Source NPZ:
- `DEX_DATA/fast_npz/bio_usdc_fresh_paper_live_fee_replay_v2.npz`

Runner:
- `dex_platform/scripts/paper_live_virtual_lp_v1.py`

Observed replay summary:
- `rows: 1`
- `decision_log_rows: 10`
- `last_action: idle`
- `last_active_route: idle`
- `strict_pass: false`
- `return_total_pct: 0.0`
- `mdd_total_pct: 0.0`
- `reason_top: idle-bio-macro-downtrend-kill`

Interpretation:
- The refreshed BIO/USDC NPZ is being consumed by the paper-live runner.
- The logged virtual-LP state is coherent and stable: all ten decisions are idle.
- Nine of the ten decisions are attributed to the macro downtrend kill and one to warmup.
- The result explains the absence of exposure: the regime gate is suppressing the route, so the replay does not yet demonstrate a non-idle LP path.

Conclusion:
- Fresh-data logging evidence is present.
- Comparison evidence is present.
- A non-idle paper-live probe is still needed before this can count as full S2 proof.