#24 — FIFO-Trends body filter

Hypothesis — what it tests

Does a minimum candle body-fraction filter on the entry bar improve FIFO-Trends signal quality?

Description

A hybrid experiment testing whether requiring the entry bar to have a directional candle body (body ≥ 30/50/70% of range, aligned with the trade) would filter out doji/indecision entries and improve FIFO-Trends. The idea was borrowed from the Zone Recovery body=0.5 filter, which had been decisive for GBP_USD ZR. Adding the body_frac dimension expanded the sweep to 10,800 configs, run through the full 3-stage pipeline on the four best pairs. The result was decisively negative: at every non-zero threshold, zero configs even passed the in-sample walk-forward stage, versus 171–186 OOS winners for the unfiltered baseline. Diagnosis: the P&F column-reversal + confirmation entry is already a momentum filter, so the candle-level check is redundant and only collapses entry frequency and delays fills; ZR's counter-trend semantics (where the body filter blocks momentum) are the opposite of trend-following's needs. A side finding — r=2 configs topping the expanded rankings — was explicitly not deployed pending independent validation.

Key result

Indicators

Algorithms

Code