← Methodology

Validation Benchmark

Phase B universe criterion: walk-forward validation on the top 20 crypto assets by market cap. A strategy passes when at least one asset shows statistically significant excess return (bootstrap p < 0.05) vs a naive buy-and-hold baseline — with Bonferroni correction across model families documented on the methodology page.

Universe evaluated
20
Top 20 crypto by market cap
Significant at 95%
1
Excess return bootstrap p < 0.05
Phase B gate
PASS
Requires ≥1 significant asset
Walk-forward setup
60/14
7-step horizon · linear model

Per-asset validation

Sorted by excess-return significance, then directional accuracy. Click a symbol for the full coin report. Export CSV ↓

AssetDir. accuracyp (vs 50%)Excess returnp (excess)Sig.
HYPE
Hyperliquid
35.6%0.053+0.00%0.035
USDT
Tether
100.0%<0.001
USDC
USDC
100.0%<0.001
USDS
USDS
95.6%<0.001
CC
Canton
62.2%0.101+0.00%1.000
ETH
Ethereum
60.0%0.180+0.00%1.000
BTC
Bitcoin
53.3%0.655+0.00%0.079
TRX
TRON
53.3%0.655+0.00%1.000
XRP
XRP
46.7%0.655+0.00%0.472
SOL
Solana
46.7%0.655+0.00%0.063
DOGE
Dogecoin
44.4%0.456+0.00%0.357
ZEC
Zcash
44.4%0.456+0.00%0.414
XLM
Stellar
42.2%0.297+0.00%1.000
LINK
Chainlink
42.2%0.297+0.00%1.000
RAIN
Rain
37.8%0.101+0.00%1.000
WBT
WhiteBIT Coin
35.6%0.053+0.00%0.328
LEO
LEO Token
33.3%0.025
BNB
BNB
31.1%0.011+0.00%1.000
XMR
Monero
31.1%0.011+0.00%0.338
FIGR_HELOC
Figure Heloc
17.8%<0.001

Confidence calibration — BTC

Reliability diagram: forecast confidence bins vs realized directional hit rate (7-step horizon, 5 bins). Perfect calibration follows the dashed diagonal.

2040%4060%6080%
Hit rate Mean confidence

Cached ~5 min. Not investment advice — see whitepaper for limitations.