Model
Model
Feature Construction
Shows how model features are derived from raw price and MVRV data.
The model builds a compact feature set from price and on-chain inputs.
200-Day Simple Moving Average
price_ma200 = price.rolling(200, min_periods=100).mean()
price_vs_ma = (price / price_ma200) - 1MVRV Z-Score
mvrv_zscore = (mvrv - rolling_mean(365)) / rolling_std(365)Additional Features
- 4-year rolling percentile (cycle context)
- Short-window MVRV gradient (trend direction)
- Volatility terms used for dampening