LAB #4152 NEUTRAL: LEDGER BENCH: Apply filter: Session = LONDON — improves win rate from 32.5% to 38.1% (+5.5pp, n=21) — E8 Intelligence Research

@misc{caldin20264152,
  title = {{LAB #4152 NEUTRAL: LEDGER BENCH: Apply filter: Session = LONDON — improves win rate from 32.5% to 38.1% (+5.5pp, n=21) — E8 Intelligence Research}},
  author = {Andrew Stewart Caldin},
  howpublished = {Zenodo},
  year = {2026},
  month = oct,
  doi = {10.5281/zenodo.23431154},
  url = {https://zenodo.org/records/23431154},
  note = {CC BY 4.0},
}

TY  - UNPB
TI  - LAB #4152 NEUTRAL: LEDGER BENCH: Apply filter: Session = LONDON — improves win rate from 32.5% to 38.1% (+5.5pp, n=21) — E8 Intelligence Research
AU  - Andrew Stewart Caldin
PY  - 2026
DA  - 2026/10/17
PB  - Zenodo
DO  - 10.5281/zenodo.23431154
UR  - https://zenodo.org/records/23431154
AB  - IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = LONDON — improves win rate from 32.5% to 38.1% (+5.5pp, n=21)  SAME-WINDOW EFFECT: Applying the LONDON session filter to the live window above would have excluded the two wkjudas trades from 2026-09-01 (GBPUSD LONG and USDJPY SHORT, both losses) and the AUDUSD SHORT via rizzy7m (loss), while retaining the NZDUSD SHORT win and the GBPUSD SHORT win. That removes 3 losses and keeps 2 wins, shifting the live window from 2W/5L (28.6%) to 2W/2L (50%) — a +21.4pp improvement, though on a tiny n=4. PnL delta is positive but dominated by avoiding the three losing trades, not by adding new winners.  OVER-TIME PROJECTION: Over 3-6 months, the LONDON filter would likely raise win rate modestly (as backtest suggests) but the real risk is overfitting to a 21-trade sample — the +5.5pp could be noise, especially since LONDON sessions often have higher volatility and wider spreads, which may degrade risk-rew  Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com
ER  -