The Kelly criterion
The mathematically optimal bet size to maximise long-run geometric growth.
John Kelly, 1956. The formula: bet a fraction of your capital equal to your edge divided by your odds. If you have a 60% chance of winning a coin flip at even money, you should bet 20% of your bankroll.
Bet too big and you risk ruin (which kills the compounding). Bet too small and you leave growth on the table. Kelly maximises the geometric growth rate of your capital over time.
For operators, the discipline is to size every bet (capital allocation, time allocation, attention allocation) according to your honest edge. Most operators systematically over-bet on low-conviction ideas and under-bet on high-conviction ones. Kelly is the formal antidote.
Examples in the wild
Portfolio company allocation by VCs is implicitly a Kelly problem. Bet too much on any single position and the fund risks ruin. Bet too little and the winners can't pay for the losers.
Buffett has occasionally said he's bet up to 50% of his capital on single positions when his conviction was high. The Kelly math allows this when the edge is real.
Career bets follow Kelly logic. The all-in version (quit your job to start a company) is only Kelly-optimal when your conviction in the new path is very high.
The Kelly criterion is one of the mental models we apply through real cases inside the Pareto MBA — a part-time program for professionals who want to think clearly about business.