Bankroll Management: The Kelly Criterion Applied

<h2>Beyond the Rule of Thumb</h2> <p>Most bankroll advice in poker circles boils down to a rule of thumb: keep 20–50 buy-ins for cash games, 100 buy-ins for tournaments. These thresholds have intuitive appeal and reasonable empirical support, but they do not explain why or account for the specific win rate, variance, and risk tolerance of each individual player. The Kelly Criterion provides a mathematically rigorous framework for thinking about optimal stake selection relative to bankroll.</p> <h2>Kelly Criterion: The Core Idea</h2> <p>John Kelly Jr. developed his criterion in 1956 in the context of information theory, but its application to gambling and investment is direct. The Kelly formula optimizes the growth rate of your bankroll over repeated bets: bet the fraction of your bankroll equal to your edge divided by the odds you are receiving.</p> <p>In a simple coin-flip analogy: if you have a 55% edge flipping a biased coin at even money, Kelly tells you to bet 10% of your bankroll on each flip (edge = 0.55 – 0.45 = 0.10, odds = 1, Kelly fraction = 0.10/1 = 10%). Betting more than Kelly grows your bankroll more slowly in the long run and increases ruin risk. Betting less than Kelly also grows more slowly but with less ruin risk.</p> <h2>Applying Kelly to Poker Cash Games</h2> <p>Poker is more complex than a coin flip — variance is continuous rather than binary — but the principles transfer. Suppose you have an estimated win rate of 5 big blinds per 100 hands with a standard deviation of 100 big blinds per 100 hands. Your effective Kelly fraction works out to approximately win rate divided by variance: 5/100² = 0.05%. In dollar terms for a $1/$2 game where 1 big blind = $2, you are putting $2 per 100 hands at risk in net expectation with $200 standard deviation per 100 hands.</p> <p>The practical takeaway: full Kelly at poker stakes would suggest an extremely large bankroll relative to the buy-in — often 200+ buy-ins — to be theoretically optimal for long-run growth. This is why most professionals adopt a fractional Kelly approach, accepting lower but still meaningful bankroll requirements by accepting a somewhat higher ruin risk in exchange for reduced capital requirements.</p> <h2>Conservative vs. Aggressive Bankroll Sizing</h2> <p>Half-Kelly — using half the mathematically optimal bet size — reduces ruin risk dramatically while surrendering only modest growth rate. For poker, this translates to: if Kelly suggests 200 buy-ins is optimal for a specific win rate and variance profile, half-Kelly means 100 buy-ins. Most professional cash game players operate somewhere in the 30–80 buy-in range, representing aggressive to conservative fractional Kelly sizing.</p> <p>The choice between aggressive and conservative sizing depends on factors beyond mathematics:</p> <ul> <li><strong>Income dependence:</strong> If poker is your only income, err conservative. Ruin is catastrophic when there is no employment backup.</li> <li><strong>Win rate certainty:</strong> The Kelly formula depends on accurately knowing your edge. Most players overestimate their win rate. When uncertain about your true edge, bias toward conservative bankroll sizing.</li> <li><strong>Game quality stability:</strong> If you have reliable access to soft games, you can be more aggressive. If game quality is variable or access uncertain, add buffer.</li> </ul> <h2>Shot-Taking: Calculated Excursions Up in Stakes</h2> <p>Shot-taking is the practice of temporarily playing higher stakes than your bankroll strictly supports, with a predetermined loss limit triggering a return to your normal stake. Done correctly, shot-taking is a rational expression of Kelly logic — you are betting more when you identify a favorable opportunity (an unusually soft game at a higher stake) and cutting the shot when the opportunity cost exceeds the expected value.</p> <p>A disciplined shot-taking framework: define your shot as a maximum of 2–5 buy-ins at the higher stake. If you lose those buy-ins, return to your normal game without exception. If you win and build your bankroll to the threshold supporting the higher stake sustainably, you have successfully moved up. If you deviate from this framework — adding more buy-ins to the shot after losses, telling yourself the game is too good to leave — you have ceased taking a shot and begun gambling with your bankroll.</p> <p>The Kelly Criterion does not guarantee against downswings. No mathematical framework does. What it guarantees — given accurate inputs — is that you are optimizing the long-run growth of your bankroll within your chosen risk tolerance. For poker players committed to treating the game as a profession, this is precisely the framework that transforms bankroll management from superstition into science.</p>