Track Record:
Transparent by Default

Methodology

Fred uses quantitative models trained on historical market data. All predictions are logged at the time they are made, before outcomes are known.

• Regime Model — 6-signal composite: BTC z-score, momentum (SMA20 vs SMA50), Fear & Greed, stablecoin supply trend, volume trend, prediction market sentiment
• ML Regime Predictor — Gradient Boosted classifier predicting regime 7 days forward
• Sector Rotation — GBM model predicting sector outperformance probability over 30-day window
• Accuracy — Measured by comparing predicted regime/sector to actual outcome on the target date

Model accuracy is self-reported from backtesting. Live accuracy tracking begins when sufficient forward predictions resolve.