2,055 bars 2021-01-01 → 2026-08-17. Rule: dual w200 rw60 (long when
close > SMA200 AND ROC60 > 0; state at close[t], fill at open[t+1]).
Headline (maker 30 bps RT)
| metric | BTC dual w200 rw60 | ETH dual w200 rw60 (phase 13) |
|---|---|---|
| full net | +12,010 bps (+143.3%) | +15,799 bps (+258.9%) |
| trades | 31 | 27 |
| max DD | −21.1% | −17.4% |
| cal half (2021–23) | +146 bps (~flat) | +4,860 bps |
| val half (2024–26) | +9,270 bps | +10,938 bps |
| taker (50 bps RT) | +11,328 bps | +15,193 bps |
| B&H same window | +114.8%, maxDD −76.7% | — |
Yearly (net bps, maker)
| year | strategy | trades | B&H |
|---|---|---|---|
| 2021 | −2,521 | ~9 | +6,106 |
| 2022 | 0 | 0 | −6,408 |
| 2023 | +2,667 | — | +15,509 |
| 2024 | +7,043 | — | +11,912 |
| 2025 | +4,821 | — | −533 |
| 2026 YTD | 0 | 0 | −2,820 |
Robustness
All neighbor windows positive on the val half: dual w100 +9,130 / w150 +9,140 / w200 +9,270 / w300 +10,008; ma w100–300 likewise (w300 ma: +14,167 full). Not a single-cell artifact.
Verdict
PASS with a caveat. Clears the trend-overlay gates: net positive full history, val half strongly positive, taker-survivable, drawdown shield real (−21% vs −77% B&H; dodged 2022 entirely). The caveat is the cal half: 2021's post-peak chop cost −25.2% and leaves cal ≈ flat — weaker than ETH's cleanly positive cal. Edge is concentrated in 2023–2025.
Deployment notes:
- Phase 13 §5.4 already showed ETH+BTC 50/50 within one account underperforms ETH alone (+176.1% vs +258.9%) — do not blend capitals. If deployed, BTC is a separate sleeve/account decision.
- Gate correlation is high by construction (BTC ROC60 already gates ETH entries), but yearly profiles differ (2021: ETH +35% vs BTC −25%), so the sleeves are not redundant.
- Same operational blocker as gate60: paper account buying power / R4.