BTC

Analysis complete. All quoted figures verified against source data.

Verdict: WAIT — Score 28/100 · 0/3 ENTER gates

select-strategy("momentum-breakout") → success:true, selectedAt:2026-09-11T05:30:25.265Z

Signal bar (last closed 1h, 04:00): C 77,073.6 · V 1,348.6 · RSI 43.56 · live mark 77,103.6

Why not ENTER

Gate	Result
Score ≥70	✗ 28
Momentum rising (RSI>50 and rising)	✗ rising hard (+5.27, slope5 +1.14/bar) but RSI 43.56 < 50
Volume >20-avg	✗ 0.53× — only 1 of 8 bounce bars cleared avg20, and that was the 23:00 down bar
HTF aligned	✗ 4h RSI 32.7 falling, price −1.57% under 4h EMA20
Key level broken	✗ FVG bottom 77,210.7 untouched (bounce high 77,165.5), 0.28 ATR short

Score 15/40 technical + 13/35 momentum. Sentiment and risk buckets are structural data gaps (packages/shared absent from this image, so risk-manager.ts cannot load — not a transient failure). Sensitivity: crediting both buckets fully still yields only 53/100, below the 60 CONFIRM line. Conclusion robust.

Three findings worth your attention

1. The trendline break is a false positive I nearly scored. Price did cross the descending trendline — but at the 02:00 close, two bars before the signal, so it's stale. That line is drawn from only two same-evening highs (−65 pt/hr) and sits 0.35% below price. Per the strategy's own priority ordering (FVG > trendline > resistance), the FVG at 77,211–77,651 is the governing level, and the bounce peaked 45 pts below it and is now fading. That's rejection under the level, not a break of it. Scored 5/16.

2. ⚠ Strategy-integrity flag. I ran the strategy-true long breakout as a control over 7,999 closed bars: negative expectancy at every horizon (n=291, −0.23%/4h, 31% win) — and the high-volume variant is negative too (n=240, −0.23%/4h). So even if all three gates flip green, this setup has no edge on BTC 1h over the past ~11 months at 20 bps. The edge in-sample is entirely with breakdown shorts (n=316, +0.22%/4h, 57% win), which is not currently triggering. Recommend a regime filter/re-derivation before any long is taken on a green signal.

3. Correction to the 04:30 digest. It reported 4h stored volume as corrupt (ratio 0.0117) and forced a rebuild from 1h. That applies to the local parquet store, not the Bybit API. Live 4h vs Σ(1h) across 250 closed bars: volume ratio median 1.0000, zero close/low deviations. 5m/15m are likewise clean this session (0 zero-volume bars). The prior run's conclusions still hold (rebuild-from-1h is a valid substitute), but the "corrupt feed" alarm shouldn't propagate — the API is fine.

Analog C matches the current state exactly, last firing on this very bar (04:00), n=219: LONG −0.33%/4h at 35% win / fade +0.33% at 65% win. Analog D is the trap specifically: 94% of such bars reclaim EMA20 within 24h but only 52% take out the prior 20-bar high. Derivatives corroborate — OI +0.75%/5h is a partial rebuild (still −1.80% vs 8h ago), funding ~neutral with no squeeze fuel, and rising retail long-share (59.69%) into a bearish stack.

Not acting on a short, despite the favorable fade expectancy — that's counter-momentum and belongs to contrarian/mean-reversion, not this mandate.

Watch plan: long becomes scoreable only on a closed 1h >77,210.7 then >77,514.8 with vol >2,525 BTC and RSI >50. Bounce invalidates on a close <76,842.4; short resumes <76,394.1 → weekly low 76,257.4. Next scan 06:30 UTC, or immediately on either trigger.

Digest: apps/website/digests/2026-09-11T0530Z-BTC-1h-momentum-breakout.md