LAB #1217 HARMFUL: LEDGER BENCH: Apply filter: Phase distance CLOSE (2-5deg) — improves win rate from 31.5% to 31.5% (+ — E8 Intelligence Research
Year: 2026
Type: preprint
Abstract: IDEA: Adjustment made to the live model — verify against the backtests: Apply filter: Phase distance CLOSE (2-5deg) — improves win rate from 31.5% to 31.5% (+0.0pp, n=111) SAME-WINDOW EFFECT: Applying the Phase distance CLOSE (2-5deg) filter to the current live trade window would have no effect, as the backtest shows zero win-rate improvement (+0.0pp) across 111 samples. None of the currently open or closed trades would have been filtered differently, since the filter does not alter outcomes. The PnL delta is zero — no trades are added or removed, and no win/loss status changes. OVER-TIME PROJECTION: Over 3-6 months, this filter adds no predictive value — the win rate remains identical at 31.5%. The risk is that it introduces unnecessary computational overhead and potential overfitting to noise, as a zero-delta filter on a small sample (n=111) is statistically indistinguishable from random variation. It may also create false confidence in a "refinement" that does nothing, leading to Author: Andrew Stewart Caldin, Independent Researcher, UK. Part of the E8 Intelligence Research series. Platform: e8intelligence.com
Author: Caldin, Andrew Stewart
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7203884653
Cites:
Cited by:
Related to: