Kelly Formula

Q: What is the Kelly formula?

A: The Kelly formula is a bankroll management method. It calculates how much capital should be committed when a probability edge exists. It does not answer only "should I buy?" It answers the next question: if this selection has positive expectation, how large should the position be?

Q: Why is it called the Kelly formula?

A: In 1956, John Kelly of Bell Labs studied the relationship between information advantage and optimal stake size while working from Claude Shannon's information theory. The original academic context was communication over noisy channels, with gambling examples used to illustrate the idea. Later, Edward Thorp applied the logic first to blackjack and then to securities arbitrage and derivatives trading. As quantitative trading and hedge funds developed, the Kelly formula became widely used as a position-sizing tool. Kelly himself was not writing a stock-investment manual.

Q: Why is the Kelly formula suitable for football quant research?

A: Football quant strategies repeatedly face the same problem: the model sees an edge, but how much should be staked? If the stake is too small, the edge is hard to express. If the stake is too large, a losing streak can destroy the capital curve. The Kelly formula provides a framework that connects probability, odds, and bankroll.

Q: What are the basic inputs?

A: The Kelly formula needs at least two inputs. The first is the true probability, meaning the probability you estimate for an outcome. The second is the odds, meaning the market price. If the true probability estimate is wrong, the staking fraction calculated by Kelly will also be distorted.

Q: Can the Kelly formula guarantee profit?

A: No. The Kelly formula only helps control capital growth and drawdown when the probability estimate is correct and the strategy is repeated over the long run. It cannot repair a wrong probability model and cannot remove short-term volatility. If the quantitative strategy has no edge, Kelly sizing will not turn it into a valid strategy.

Q: Why do real bettors often use half Kelly or less?

A: Because true probability is rarely estimated with precision. Football matches also have limited samples, odds movement, team news, and model error. Half Kelly or lower fractions reduce the capital volatility caused by estimation mistakes and make the strategy easier to execute for a long time.

Q: How is the Kelly formula calculated?

A: In the simple betting version:

\[ f^*=\frac{bp-q}{b} \]

Here, \(f^*\) is the optimal staking fraction, \(b\) is the net odds, \(p\) is the win probability, and \(q=1-p\) is the loss probability.

This simple form applies to a binary bet: winning earns \(b\) times the stake, and losing loses the full stake.

Q: Where does the Kelly formula sit in this site's knowledge graph?

A: The Kelly formula is the bankroll management layer. It connects upward to odds and quantitative strategy, sideways to backtesting, and downward to concrete staking examples and Python implementations.