Window since 2026-09-01 (UTC), 1 bot(s).
| Bot | Date | Live | Sim | Match | Slippage (bps) | Fees |
|---|
Summary: no comparable decisions Mismatches: 0
No slippage records in window (all OBSERVE or no fills).
Reading
✗ MISMATCHmeans live and replayed frozen decide disagree on a completed-bar date -> investigate data feed or warm-up divergence.- Large or systematic slippage vs paper assumption suggests sizing or maker logic needs tuning (P1b/P1c).
Flagship (region-A) — live trades vs traded cache
| Date | Sym | Action | Reason | Live px | Cache combined | ml_active | Cache close | Verdict |
|---|---|---|---|---|---|---|---|---|
| 2026-09-04 | AVGO | buy | entry | 358.25 | +0.216 | True | 357.90 | OK |
| 2026-09-04 | DXCM | sell | signal_exit | 87.81 | -0.220 | False | 87.90 | MISMATCH (exit w/o flip) |
| 2026-09-04 | MSCI | buy | entry | 573.58 | +0.225 | True | 573.01 | OK |
| 2026-09-04 | VRSK | buy | entry | 185.98 | +0.165 | False | 185.79 | MISMATCH (entry w/o signal) |
| 2026-09-04 | WBD | buy | entry | 28.28 | +0.300 | True | 28.25 | OK |
Flagship signal mismatches: 2/5.
Accounting (broker truth vs local model)
Last row 2026-09-08: broker equity=$31,665.11 cash=$0.00 peak=$31,665.11
Pre-2026-09-10 equity.csv rows are local-model (BP era, inflated); post-fix rows are broker truth.
Reading
MISMATCH (entry w/o signal)means the live trade's cache bar had combined<=0.05 or ml_active=False -> investigate TOP_K gating or stale cache.MISMATCH (exit w/o flip)on signal_exit without combined<-0.25 -> investigate exit wiring.- Accounting WARNINGs point at leverage/sync bugs, not alpha.