RollingVariance Node
Rolling price volatility measurement
Overview
Rolling Variance measures the spread of returns around mean in rolling windows. It's the squared standard deviation. Rising rolling variance = market becoming more volatile. Falling rolling variance = market stabilizing/consolidating. Variance is the input to most risk models and position sizing algorithms.
Unlike single historical volatility, rolling variance shows volatility trend. Markets with rising volatility need wider stops and smaller positions. Markets with falling volatility can accommodate larger positions. This fundamental metric drives nearly all active trading decisions.
Formula & Calculation
StdDev = sqrt(Variance)
Annualized: Var_annual = Var_daily × 252
Parameters
| Parameter | Default | Description |
|---|---|---|
| lookback | 20 | Rolling window |
| ddof | 1 | Sample variance (ddof=1) |
Common Use Cases
1. Position Sizing
Position = Target Risk / sqrt(Rolling Variance). High variance = smaller positions. Low variance = larger positions. Dynamically adapt to volatility regime.
2. Stop Loss Setting
Stops = 2 × sqrt(Rolling Variance). High volatility = wider stops. Low volatility = tighter stops. Always proportional to realized volatility.
3. Leverage Adjustment
Leverage = Base × Normal_Var / Current_Var. When variance spikes, reduce leverage automatically. Dynamic de-leveraging protects during crisis.
4. Risk Modeling
Rolling variance is input to VaR, CVaR, Sharpe, and most risk calculations. Track its trend: rising variance = harder period ahead.
Advantages & Limitations
Advantages
- Fundamental metric
- Directly usable for sizing
- Easy to calculate
- Well understood
Limitations
- Assumes normality
- Backward looking
- Can't forecast jumps
- Lagged indicator