BTC

BTC/USDT · 1h · MOMENTUM BREAKOUT — WAIT (score 33/100)

Strategy selected: momentum-breakout (scheduled scan). Digest persisted to digests/2026-09-11T0700Z-BTC-1h-momentum-breakout.md. All market data pulled live from Bybit linear (10,000-bar 1h sample, zero integrity defects) with Binance cross-check (0.062% venue spread).

Gates — 0/3 pass

Gate	Result
Score ≥ 70	✗ 33
RSI > 50 and rising	✗ rising (slope5 +3.00) but RSI 47.90 < 50, and decelerating: +3.17 → +1.16/bar
Volume > 20-avg	✗ 0.55× (1,384 vs 2,529). Every one of the last 4 up-closes sub-average
HTF aligned	✗ 4h RSI 32.7 falling, price < 4h EMA20; 1w bear stack; last 4 swing highs strictly descending
Key level broken	✗ still INSIDE bear FVG 77,210.7–77,650.9; trigger 0.93 ATR away

Score: Technical 17/40 · Momentum 13/35 · Sentiment 3/10 · Risk 0/15 (gap)

Technical did improve (15→17) on the first 1h close above EMA20. The gap is proven immaterial: full marks to Risk still caps at 48/100 — CONFIRM is unreachable.

The finding that decides it

Price advanced and closed above the 1h EMA20 — but the same bar rejected the weekly EMA50: the 06:10 spike hit 77,395.2, clearing the weekly EMA50 (77,383.9) on the bounce's largest 5m volume, then collapsed to close 130 points below the high. Supply won the first real test. Stoch J rolled 102.6 → 92.2, OI fell −0.65% into the strength (rally on position closure, no fresh risk), and retail long share rose to 59.8%.

Base rates on 9,999 bars (20bps cost) — the pattern is net loss-making even when it fires:

• Strategy-true long breakout (n=382): −0.23% to −0.27%, win 32–44%
• Live-state analog, 4 up-closes on no volume in a bear stack (n=62): −0.50% to −0.68%, win 26–34% — worst bucket tested
• Trigger-first vs breakdown-first within 24h from the live state: 31% vs 62%
• The only positive expectancy in the sample is the strategy-true short (n=387, +0.27–0.41%, 56–59%) — and it is not armed
• The volume gate earns its keep: it adds +0.19% and +10pp win rate versus the current ungated state

The short side was explicitly checked and rejected too (no 20-bar low break, RSI rising, no volume). Its planned R:R also fails 2:1 — prev-week low 76,257.4 sits at 1.36R.

Armed triggers (no orders placed)

• LONG: closed 1h bar > 77,650.9 AND vol > 2,529 AND RSI > 50 rising → entry 77,651 / stop 76,869 / 2R 79,215. If filled, halve size and take profit into 78,499–78,630; the confirmed-trigger subset graded to −0.089R, 36% win.
• SHORT: close < 76,868.9 on >1.0× volume — but not actionable at that stop; needs 76,257.4 to break with volume.

Honest data caveats: risk-manager.ts cannot load in this runtime (ERR_UNSUPPORTED_TYPESCRIPT_SYNTAX; packages/shared absent) → Risk bucket unscored, decision proven insensitive. Bybit taker buy/sell ratio genuinely unavailable (404) — flagged, not substituted. I recovered the orderbook via the correct /v5/market/orderbook path (the 404'ing /orderbook/depth used previously), but top-50 imbalance swung 1.87→5.53 within six seconds while 200-level depth read 1.24 — so I discounted it as noise rather than treat it as confirmation, and it is not load-bearing anywhere in this call.