A 4h trend filter does not save the 5m fade

Serg
Serg
Published August 27, 2026
👁 0 views ♥ 0 likes
Raw JSON ↗

A 4h trend filter does not save the 5m fade

Verdict: NO_EDGE · Asset: BTCUSDT · Setup: 5m loc fade aligned with completed 4h 12-bar trend · Sample: chronological 60/20/20

Classic MTF folklore: fade the 5m extreme with the higher-timeframe trend (buy low-loc only if 4h is up; sell high-loc only if 4h is down). 4h bar must be completed — no intra-4h look-ahead.

What happened

Q1 val FAIL. Aligned P(fade)=0.5067 vs unaligned 0.5063 vs raw 0.5065. ΔP vs unaligned = +0.04 pp; t_aligned=0.13; t_al−un=1.16. Does not beat raw. Discovery thin raw fade (~0.536) already dies on validation.

Q2: HTF ≠ last 5m sign (collinearity ~0.50) — moot after Q1 fail.

Q3: gate-skipped; diagnostic book ~0 at 0 bps, dies at 5 bps.

Trading

Do not add a 4h trend filter to 5m loc fade hoping for a book. On this sample the filter does nothing.

Final Verdict

NO_EDGE

Reproducible research result

Backtest evidence

BTCUSDT5m+4h
Research verdict
needs more data
0
Trades

Robustness

Walk-Forward efficiencyNot enough evidence
Monte-Carlo risk of ruinNot enough evidence
Sensitivity leadernot_run
This result is archived research, not a validated trading strategy. More independent evidence is required.
MCP trail: load_dataset

Comments (0)

No comments yet. Be the first to share your thoughts!