keeks
Kelly-criterion bet sizing and bankroll allocation in Python
🧩 Build it Free 02 in Risk & sizing 🆕 NEW TODAY
6 0 copies · 7 days MITupdated 8 Sep '26
Overview
Keeks calculates a bankroll fraction to stake from nine bet-sizing strategies, including Kelly Criterion, then simulates the bankroll path over repeated trials.
- Fordevelopers modeling repeated binary-outcome bets or trades who want a sizing rule and a bankroll simulator, not a one-off price.
- NeedsPython 3.10-3.14; you supply your own win probability, payoff, loss, and transaction cost — no data source or account.
- Runsa pip-installed Python library — a strategy object's evaluate() call, run through a BankRoll and a repeated-bet simulator.
- Limitthe fraction it returns answers a repeated-bet question — sizing a single, one-time trade off it is the wrong reading; use find_indifference_price for a one-time gamble instead.
Vibe it
Paste this into your AI agent. Write the first line, or take a suggestion; the copy carries your words.
Source: https://github.com/wdm0006/keeks — clone it and read its README and examples before writing anything; its documented setup beats what you remember. Stack: a Python library — a BaseStrategy subclass (Kelly Criterion and eight others) that returns a bankroll fraction, plus a BankRoll object and a RepeatedBinarySimulator. Needs: Python 3.10-3.14, and your own estimated win probability, payoff, loss, and transaction cost — no data source or account. Start: `pip install keeks`, then run the KellyCriterion example from the README against my own probability and payoff before wiring in a BankRoll simulation. Limit: the fraction answers a repeated-bet question — sizing a single, one-time trade off it is the wrong reading; use find_indifference_price for a one-time gamble instead. Done means: you show me the bankroll fraction and dollar amount for my own inputs, and the strategy code you wrote. Before you start, ask me: my estimated win probability, payoff, loss, and transaction cost, and whether I want a single fraction or a multi-trial bankroll simulation.
