#67 — Structural fade at hourly S/R

Hypothesis — what it tests

Whether giving a contrarian entry its direction from market structure — fading the first approach to an hourly swing support/resistance level on heavy volume — yields a directional edge that survives the retail spread.

Description

The culmination of a long random-entry → spread-band → exit-investigation research arc on S5 data with real per-bar spread. Random coin-flip entries first established the floor (no exit or spread band rescues a coin flip; deep-loss tails peak at the median spread band), then giving the entry direction from structure — fading the last confirmed H1 swing high (resistance) or low (support), gated by high tick-volume and a calm spread band, with ATR-scaled distance/trail/TP — produced a gross directional edge of +1.089 pips/trade with Monte Carlo P=0.0000, positive on all 12 pairs and in all 3 walk-forward thirds: the cleanest broad positive-direction result in the project. Net of the real 1.3–3.9 pip spread it is −0.890 pips/trade, negative on all 12 pairs. Follow-ups showed level clustering/strength adds nothing, the edge is timeframe-invariant, an MFE oracle would clear spread easily (+6.44p mean) but eight exit families all converge to only ~+1.3–1.5p capturable drift (the rest is uncatchable wiggle), and the optimal hold is ~2h. The gap to the deployable first-touch H4 cousin (entry #68) is selectivity, not exit cleverness.

Key result

Indicators

Algorithms

Code