Pre-registered proposal: reports/research_proposal_carry_basis_overlay.md. Defaults: look=168h, carry|z|>=1.5, basis|z|>=2.0, multipliers {0.5, 1.0, 1.25}. Window: 2024-06-04 .. 2026-07-31 (788 days, 285 BTC-gate60 long days) | Hyperliquid BTC funding cache for CARRY+BASIS; BTC daily bars from validation.common; _single_coin_decide on BTC gated by BTC ROC60, validated in reports/universe_generalization.md used to compute the BTC-gate60 target.
Diagnostics: corr(carry_z, basis_z) = +0.721 | fraction of long days with both signals stretched long (0.5x bucket): 0.4% | fraction with both signals stretched short (1.25x bucket): 1.9%.
| Variant | Days | LongDays | Ret | Sharpe | PF | MaxDD |
|---|
| baseline (gate60, no overlay) | 788 | 285 | +39.20% | 0.75 | 1.20 | -17.58% |
| primary overlay (0.5/1.0/1.25) | 788 | 285 | +38.76% | 0.74 | 1.20 | -17.76% |
| Bucket | Days | Share |
|---|
| 1.0x (no signal) | 226 | 79.3% |
| 1.0x (mixed) | 41 | 14.4% |
| 1.25x (both stretched short) | 15 | 5.3% |
| 0.5x (both stretched long) | 3 | 1.1% |
Bootstrap p(overlay beats baseline, full window): 0.517
Bootstrap p(IS < 2025-01-01): 0.689
Bootstrap p(OOS >= 2025-01-01): 0.393
| Segment | Variant | LongDays | Ret | Sharpe | PF |
|---|
| IS | baseline | 100 | +30.39% | 1.58 | 1.41 |
| IS | primary overlay | 100 | +29.34% | 1.51 | 1.39 |
| OOS | baseline | 185 | +6.76% | 0.31 | 1.08 |
| OOS | primary overlay | 185 | +7.28% | 0.32 | 1.08 |
| look (h) | carry z_enter | basis z_enter | LongDays | Ret | Sharpe | PF | MaxDD |
|---|
| 84 | 1.0 | 1.0 | 285 | +52.56% | 0.89 | 1.25 | -17.74% |
| 84 | 1.0 | 1.5 | 285 | +47.55% | 0.85 | 1.23 | -16.99% |
| 84 | 1.0 | 2.0 | 285 | +46.20% | 0.84 | 1.23 | -16.92% |
| 84 | 1.5 | 1.0 | 285 | +50.42% | 0.88 | 1.25 | -17.01% |
| 84 | 1.5 | 1.5 | 285 | +45.84% | 0.83 | 1.23 | -16.99% |
| 84 | 1.5 | 2.0 | 285 | +44.98% | 0.83 | 1.23 | -16.93% |
| 84 | 2.0 | 1.0 | 285 | +41.24% | 0.77 | 1.21 | -16.94% |
| 84 | 2.0 | 1.5 | 285 | +41.82% | 0.78 | 1.21 | -16.93% |
| 84 | 2.0 | 2.0 | 285 | +42.74% | 0.79 | 1.22 | -16.93% |
| 168 | 1.0 | 1.0 | 285 | +53.43% | 0.90 | 1.26 | -17.49% |
| 168 | 1.0 | 1.5 | 285 | +47.68% | 0.84 | 1.24 | -17.90% |
| 168 | 1.0 | 2.0 | 285 | +39.97% | 0.75 | 1.20 | -16.93% |
| 168 | 1.5 | 1.0 | 285 | +51.33% | 0.87 | 1.25 | -17.49% |
| 168 | 1.5 | 1.5 | 285 | +46.08% | 0.82 | 1.23 | -17.90% |
| 168 | 1.5 | 2.0 | 285 | +38.76% | 0.74 | 1.20 | -17.76% |
| 168 | 2.0 | 1.0 | 285 | +42.08% | 0.76 | 1.21 | -17.90% |
| 168 | 2.0 | 1.5 | 285 | +42.07% | 0.76 | 1.21 | -17.90% |
| 168 | 2.0 | 2.0 | 285 | +38.76% | 0.74 | 1.20 | -17.76% |
| 336 | 1.0 | 1.0 | 285 | +55.13% | 0.92 | 1.26 | -16.51% |
| 336 | 1.0 | 1.5 | 285 | +50.14% | 0.86 | 1.24 | -17.16% |
| 336 | 1.0 | 2.0 | 285 | +42.57% | 0.79 | 1.21 | -17.40% |
| 336 | 1.5 | 1.0 | 285 | +49.25% | 0.84 | 1.24 | -17.52% |
| 336 | 1.5 | 1.5 | 285 | +47.17% | 0.82 | 1.23 | -17.52% |
| 336 | 1.5 | 2.0 | 285 | +40.06% | 0.76 | 1.20 | -17.58% |
| 336 | 2.0 | 1.0 | 285 | +48.72% | 0.84 | 1.24 | -17.52% |
| 336 | 2.0 | 1.5 | 285 | +48.04% | 0.83 | 1.23 | -17.52% |
| 336 | 2.0 | 2.0 | 285 | +40.06% | 0.76 | 1.20 | -17.58% |
- IS baseline Sharpe: 1.58, ret +30.39%, maxDD -12.55%
- IS overlay Sharpe: 1.51, ret +29.34%, maxDD -13.38%
- OOS baseline Sharpe: 0.31, ret +6.76%, maxDD -17.58%
- OOS overlay Sharpe: 0.32, ret +7.28%, maxDD -17.76%
- Pre-registered pass: OOS Sharpe +>= 0.10 AND p < 0.05 AND maxDD not +> 5pp
- Result: INERT (2.3% of long days fire 0.5x/1.25x, p=0.393)
- Inert signal risk: if <5% of long days trigger either 0.5x or 1.25x, the overlay is a no-op and the experiment is a wash. Reported as INERT in the verdict.
- Threshold tuning: the proposal pre-registers look=168h, carry_enter=1.5, basis_enter=2.0. The robustness grid is REPORTED, not selected. Picking the IS winner would be the look-ahead bug class that bit R10-13.
- Hyperliquid only: Binance/Bybit/OKX funding are geo-blocked from this IP (451/403). Hyperliquid started in 2023; the cache is the longest available free source. Binance USDT-M funding IS in the metrics cache but is perpetual-rate only, not the premium index; the BASIS leg specifically needs premium.
- Hyperliquid is one venue: cross-venue funding arbitrage (HL vs Binance) is not what this study tests; this study asks whether HL alone is informationally useful as a sizing overlay on a slow-trend signal.
- CARRY and BASIS are correlated: corr reported above. If corr > 0.7, the composite is one-dimensional and the two-signal framing is misleading; the overlay is then a single-signal gate with two thresholds.
- Multipliers are heuristic: 0.5x and 1.25x are the natural ±50% / ±25% asymmetry, not the result of an optimization. An optimization step would re-introduce the threshold-tuning look-ahead; the honest path is to report the asymmetry you chose and let the verdict be what it is.