IPositionSizer<T>
Interfaces · AiDotNet.Finance.Interfaces
A bet/position-sizing rule: given an edge, returns the fraction of capital to allocate.
Known implementations: KellyCriterion<T>
For Beginners
A “position sizer” answers “how big should this bet be?” The default is the Kelly criterion (the growth-optimal fraction); swap in your own rule if you prefer.
How It Works
This is a customization point, not a trainable model. Trading agents and portfolio models depend on
this interface to decide how much to commit to a signal and default to the
KellyCriterion implementation, but callers can substitute
a fixed-fraction, volatility-target, or risk-budget sizer without changing the consuming model.
Methods
| Method | Summary |
|---|---|
T Continuous(T expectedReturn, T variance) |
Sizing fraction from the (Gaussian) mean and variance of returns. |
T Discrete(T winProbability, T winLossRatio) |
Sizing fraction from a discrete win probability and win/loss payoff ratio. |
T Fractional(T baseFraction, double fraction) |
Scales a base sizing fraction (e.g. |
T FromReturns(IEnumerable<T> returns, double fraction) |
Sizing fraction estimated from a realized return series. |