Classic pin bars are just close-location in a costume

Verdict: NO_EDGE · Asset/TF: BTCUSDT 1h (primary), 15m · Sample: chronological 60/20/20

Refined pin-bar (distinct from unused-wick % and from loc-only): large rejection wick (discovery P80), body/range ≤ 0.30, close on the opposite half of the range. Classic folklore = continuation after the pin.

What happened

Q1 continuation fails (wrong sign). Bullish pin ΔP=−0.0165, t=−1.18; bearish pin ΔP=+0.0249, t=+2.79 (fade-tilt). 15m same story.

Q2 costume kill. Loc extremes dominate (high-loc ΔP=−0.058 vs pin −0.016; low-loc +0.068 vs pin +0.025). Pin ≈ large-wick alone. The fade-clear is not a new object — Q2 overrides promoting it as a pin edge.

Q3 skipped for promotion. Fade curiosity positive at 0 bps, dies at 5 bps — same wall as the thin-fade family.

Trading

Do not ship pin-bar bots on BTC klines. You already tested the real pieces (loc, wick); the candle name does not add a third factor.

Final Verdict

NO_EDGE