Quant Library · Markets and staking · Any
Kelly Staking and Portfolio Kelly
Growth-optimal stakes, fractional and capped, and the joint version for exclusive outcomes.
Single-bet Kelly, the fractional and capped version any sane desk actually uses, and the concave programme for staking several outcomes of the same event at once.
Inputs
4What it takes
| Name | Type | Units | Where it comes from |
|---|---|---|---|
| probabilities | list | probability | your model |
| odds | list | decimal | your book |
| fraction, cap optional | floats | - | your risk policy; a quarter and 5% are the platform's |
| bankroll | float | currency | - |
Outputs
3What it gives back
| Name | Type | Units | Notes |
|---|---|---|---|
| stake fractions | list | fraction of bankroll | |
| expected log growth | float | nats per bet | |
| certainty equivalent | float | currency |
Method
How it works
- For a single bet, f* = (p o − 1) / (o − 1); take a fraction of it and cap it.
- 2 further steps in the licensed specification
Assumptions
- Your probability is right. It is not, which is the entire argument for the fraction.
Limits
- Kelly is optimal for log wealth over many independent repetitions. If you will place fifty bets, not fifty thousand, the variance is the thing that will decide your year.
- 1 further limits in the licensed specification
In the package
- The full specification: inputs, method, assumptions and limits
- The validation panel as measured on the day of purchase
- A reference implementation extracted from the running source
- Perpetual commercial use for one entity
Reference modules:
- 2 module(s), named in the licensed specification and shipped in the package
Read this before you buy. The platform's own walk-forward tests say the closing line forecasts football better than these models do. What they give you is a coherent price for every market and match, before the market opens, with the band and the working attached. Any model here that has been beaten by something simpler says so in its limits, and the validation panel above is generated from the platform's result files rather than typed in.