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, 145 LINK-gate60 long days) | Hyperliquid LINK funding cache for CARRY+BASIS; LINK daily bars from validation.common; _single_coin_decide on LINK gated by BTC ROC60, validated in reports/universe_generalization.md used to compute the LINK-gate60 target.
Diagnostics: corr(carry_z, basis_z) = +0.658 | fraction of long days with both signals stretched long (0.5x bucket): 0.5% | fraction with both signals stretched short (1.25x bucket): 1.5%.
| Variant | Days | LongDays | Ret | Sharpe | PF | MaxDD |
|---|
| baseline (gate60, no overlay) | 788 | 145 | +135.34% | 1.02 | 1.41 | -33.88% |
| primary overlay (0.5/1.0/1.25) | 788 | 145 | +216.07% | 1.23 | 1.55 | -31.59% |
| Bucket | Days | Share |
|---|
| 1.0x (no signal) | 111 | 76.6% |
| 1.0x (mixed) | 18 | 12.4% |
| 1.25x (both stretched short) | 12 | 8.3% |
| 0.5x (both stretched long) | 4 | 2.8% |
Bootstrap p(overlay beats baseline, full window): 0.001
Bootstrap p(IS < 2025-01-01): 0.010
Bootstrap p(OOS >= 2025-01-01): 0.015
| Segment | Variant | LongDays | Ret | Sharpe | PF |
|---|
| IS | baseline | 55 | +42.17% | 1.16 | 1.42 |
| IS | primary overlay | 55 | +75.74% | 1.55 | 1.66 |
| OOS | baseline | 90 | +65.53% | 1.01 | 1.41 |
| OOS | primary overlay | 90 | +79.85% | 1.12 | 1.47 |
| look (h) | carry z_enter | basis z_enter | LongDays | Ret | Sharpe | PF | MaxDD |
|---|
| 84 | 1.0 | 1.0 | 145 | +349.36% | 1.51 | 1.74 | -24.53% |
| 84 | 1.0 | 1.5 | 145 | +303.61% | 1.42 | 1.68 | -24.53% |
| 84 | 1.0 | 2.0 | 145 | +265.25% | 1.35 | 1.63 | -25.64% |
| 84 | 1.5 | 1.0 | 145 | +262.12% | 1.32 | 1.61 | -28.83% |
| 84 | 1.5 | 1.5 | 145 | +263.39% | 1.33 | 1.62 | -28.83% |
| 84 | 1.5 | 2.0 | 145 | +236.46% | 1.27 | 1.58 | -29.88% |
| 84 | 2.0 | 1.0 | 145 | +220.04% | 1.22 | 1.55 | -28.83% |
| 84 | 2.0 | 1.5 | 145 | +220.04% | 1.22 | 1.55 | -28.83% |
| 84 | 2.0 | 2.0 | 145 | +209.47% | 1.19 | 1.54 | -29.88% |
| 168 | 1.0 | 1.0 | 145 | +286.21% | 1.38 | 1.65 | -28.83% |
| 168 | 1.0 | 1.5 | 145 | +253.91% | 1.31 | 1.61 | -28.83% |
| 168 | 1.0 | 2.0 | 145 | +216.07% | 1.23 | 1.55 | -31.59% |
| 168 | 1.5 | 1.0 | 145 | +254.22% | 1.31 | 1.60 | -28.83% |
| 168 | 1.5 | 1.5 | 145 | +236.79% | 1.27 | 1.58 | -28.83% |
| 168 | 1.5 | 2.0 | 145 | +216.07% | 1.23 | 1.55 | -31.59% |
| 168 | 2.0 | 1.0 | 145 | +196.91% | 1.16 | 1.52 | -29.88% |
| 168 | 2.0 | 1.5 | 145 | +196.91% | 1.16 | 1.52 | -29.88% |
| 168 | 2.0 | 2.0 | 145 | +190.71% | 1.15 | 1.51 | -31.59% |
| 336 | 1.0 | 1.0 | 145 | +268.53% | 1.34 | 1.62 | -29.47% |
| 336 | 1.0 | 1.5 | 145 | +217.13% | 1.23 | 1.55 | -32.18% |
| 336 | 1.0 | 2.0 | 145 | +202.19% | 1.19 | 1.53 | -32.18% |
| 336 | 1.5 | 1.0 | 145 | +229.86% | 1.25 | 1.57 | -32.18% |
| 336 | 1.5 | 1.5 | 145 | +216.21% | 1.22 | 1.55 | -32.18% |
| 336 | 1.5 | 2.0 | 145 | +201.32% | 1.19 | 1.53 | -32.18% |
| 336 | 2.0 | 1.0 | 145 | +184.92% | 1.13 | 1.50 | -32.18% |
| 336 | 2.0 | 1.5 | 145 | +184.92% | 1.13 | 1.50 | -32.18% |
| 336 | 2.0 | 2.0 | 145 | +177.14% | 1.11 | 1.49 | -32.18% |
- IS baseline Sharpe: 1.16, ret +42.17%, maxDD -31.63%
- IS overlay Sharpe: 1.55, ret +75.74%, maxDD -29.27%
- OOS baseline Sharpe: 1.01, ret +65.53%, maxDD -20.53%
- OOS overlay Sharpe: 1.12, ret +79.85%, maxDD -20.53%
- Pre-registered pass: OOS Sharpe +>= 0.10 AND p < 0.05 AND maxDD not +> 5pp
- Result: PASS
- 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.