Figure 1  ·  The Vault Test

Three-Strategy Mean Return by Year Near-Blind Forward Validation

2023–2025 out-of-sample window  ·  51 folds (17 assets × 3 test years)  ·  All strategy parameters locked on 2019–2022 training data
0% +25% +50% +75% +100% −25% Mean Return (17-fold year) +102% −0.6% +10.5% 2023 recovery year STXUSD 2023 sleeve (+60.5%) drives result +89% +3.3% +3.8% 2024 bull cycle 6 active grid folds · most distributed −22% +0.2% +5.4% incl. ZEC ‡ +0.2% excl. ZEC 2025 bear-dominated ‡ partial-year ‡ ZECUSD 2025 — ongoing fold Data cutoff 2026-05-26. Excl. ZEC: +2.0% mean Buy-and-Hold (17-fold year mean) Two-State RXI Three-State RXI (+ sleeve) 51-fold mean: B&H +46.2% · Two-State +0.6% · Three-State +3.8% (+2.0% excl. ZECUSD 2025 ongoing fold)
Metric — all 51 folds Buy-and-Hold Two-State RXI Three-State RXI
Mean return (all 51 folds) +46.2% +0.6% +3.8%
Mean return (excl. ZECUSD 2025 ongoing) — — +2.0%
Folds with positive return 29 (56.9%) 15 (29.4%) 25 (49.0%)
Folds negative 22 (43.1%) 2 (3.9%) 2 (3.9%)
Bear-fold win rate (20 folds) — 100% (20/20) 100% (20/20)
Simulated / Backtested Performance. All values are simulated/backtested estimates under the walk-forward methodology described in the study. Parameters were locked on 2019–2022 training data before any 2023–2025 prices were observed. Residual in-sample bias exists from parabolic detector calibration on 2021 events. The ZECUSD 2025 fold is ongoing (data cutoff 2026-05-26) and the +80.7% sleeve contribution may change materially. Past simulated performance does not guarantee future results.
Yodacom Research · The Vault Test 2023–2025 · grid-blind-validation-2023-2025