LAB #4176 NEUTRAL: LEDGER BENCH: Apply filter: Session = OFF_SESSION — improves win rate from 32.6% to 42.5% (+9.9pp, n — E8 Intelligence Research
- 1Independent Researcher, United Kingdom
Abstract
IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = OFF_SESSION — improves win rate from 32.6% to 42.5% (+9.9pp, n=127)
SAME-WINDOW EFFECT: Applying the OFF_SESSION filter would have excluded the three wkjudas trades (GBPUSD LONG, USDJPY SHORT, NZDUSD SHORT) and the rizzy5m AUDUSD SHORT from 08-27, all of which were losses or open/rejected. That removes 4 of 10 trades, eliminating 3 confirmed losses and 1 open/rejected, leaving only the two rizzy5m wins and the rizzy7m loss. Expected win rate in this window would rise from 3/10 (30%) to 2/3 (66.7%), but PnL delta is ambiguous — you lose the rejected/open trade's potential and the sample shrinks to near-zero statistical power.
OVER-TIME PROJECTION: Over 3-6 months, the +9.9pp improvement on n=127 is real but fragile — it likely reflects a regime where OFF_SESSION hours (Asian/early London) had lower volatility and cleaner trends, not a structural edge. Risk: the filter may overfit to a speci
Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com
Related
- LAB #4419 NEUTRAL: LEDGER BENCH: Apply filter: Session = OFF_SESSION — improves win rate from 31.6% to 41.9% (+10.3pp, — E8 Intelligence Research
- LAB #4462 NEUTRAL: LEDGER BENCH: Apply filter: Session = LONDON_OPEN — improves win rate from 31.5% to 34.8% (+3.3pp, n — E8 Intelligence Research
- LAB #4435 NEUTRAL: LEDGER BENCH: Apply filter: USDCHF SHORT @ OFF_SESSION with 6 conf — improves win rate from 31.5% to — E8 Intelligence Research
- LAB #4494 NEUTRAL: LEDGER BENCH: Apply filter: USDCHF SHORT @ OFF_SESSION with 6 conf — improves win rate from 31.6% to — E8 Intelligence Research
- LAB #4589 NEUTRAL: LEDGER BENCH: Apply filter: Session = NY_OPEN — improves win rate from 31.6% to 38.3% (+6.7pp, n=802 — E8 Intelligence Research
BibTeX
@misc{caldin20264176,
title = {{LAB #4176 NEUTRAL: LEDGER BENCH: Apply filter: Session = OFF_SESSION — improves win rate from 32.6% to 42.5% (+9.9pp, n — E8 Intelligence Research}},
author = {Andrew Stewart Caldin},
howpublished = {Zenodo},
year = {2026},
month = oct,
doi = {10.5281/zenodo.23446975},
url = {https://zenodo.org/records/23446975},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - LAB #4176 NEUTRAL: LEDGER BENCH: Apply filter: Session = OFF_SESSION — improves win rate from 32.6% to 42.5% (+9.9pp, n — E8 Intelligence Research
AU - Andrew Stewart Caldin
PY - 2026
DA - 2026/10/18
PB - Zenodo
DO - 10.5281/zenodo.23446975
UR - https://zenodo.org/records/23446975
AB - IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Session = OFF_SESSION — improves win rate from 32.6% to 42.5% (+9.9pp, n=127) SAME-WINDOW EFFECT: Applying the OFF_SESSION filter would have excluded the three wkjudas trades (GBPUSD LONG, USDJPY SHORT, NZDUSD SHORT) and the rizzy5m AUDUSD SHORT from 08-27, all of which were losses or open/rejected. That removes 4 of 10 trades, eliminating 3 confirmed losses and 1 open/rejected, leaving only the two rizzy5m wins and the rizzy7m loss. Expected win rate in this window would rise from 3/10 (30%) to 2/3 (66.7%), but PnL delta is ambiguous — you lose the rejected/open trade's potential and the sample shrinks to near-zero statistical power. OVER-TIME PROJECTION: Over 3-6 months, the +9.9pp improvement on n=127 is real but fragile — it likely reflects a regime where OFF_SESSION hours (Asian/early London) had lower volatility and cleaner trends, not a structural edge. Risk: the filter may overfit to a speci Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com
ER -