LAB #4593 NEUTRAL: LEDGER BENCH: Apply filter: Quarterly phase = MANIPULATION — improves win rate from 31.5% to 32.3% ( — E8 Intelligence Research

@misc{caldin20264593,
  title = {{LAB #4593 NEUTRAL: LEDGER BENCH: Apply filter: Quarterly phase = MANIPULATION — improves win rate from 31.5% to 32.3% ( — E8 Intelligence Research}},
  author = {Andrew Stewart Caldin},
  howpublished = {Zenodo},
  year = {2026},
  month = oct,
  doi = {10.5281/zenodo.23587134},
  url = {https://zenodo.org/records/23587134},
  note = {CC BY 4.0},
}

TY  - UNPB
TI  - LAB #4593 NEUTRAL: LEDGER BENCH: Apply filter: Quarterly phase = MANIPULATION — improves win rate from 31.5% to 32.3% ( — E8 Intelligence Research
AU  - Andrew Stewart Caldin
PY  - 2026
DA  - 2026/10/23
PB  - Zenodo
DO  - 10.5281/zenodo.23587134
UR  - https://zenodo.org/records/23587134
AB  - IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Quarterly phase = MANIPULATION — improves win rate from 31.5% to 32.3% (+0.8pp, n=1545)  SAME-WINDOW EFFECT: Applying the MANIPULATION-phase filter would have suppressed the three REJECTED/open trades (USDCHF LONG, EURJPY SHORT, AUDUSD LONG) — all currently unresolved, so no direct win/loss change yet. Among the seven closed trades, none fall into a clearly identified MANIPULATION phase based on the provided timestamps and entries, so the filter would not have altered any closed outcomes. Net PnL delta over this window is effectively zero, but it would have reduced open exposure by three positions, lowering risk without changing realized results. The +0.8pp backtest gain is not visible here due to the small sample and phase ambiguity.  OVER-TIME PROJECTION: Over 3–6 months, the filter could plausibly deliver a modest win-rate improvement if the backtest's phase classification is stable and not overfit  Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com
ER  -