Kelly Love And Money

A Practical Guide to the Kelly Love And Money Framework

The Kelly Criterion is one of those formulas people hear about in passing and immediately misapply. The version most online guides write about works for casino blackjack or sports betting with known probabilities. The Kelly Love And Money approach extends the same logic into personal finance and relationship decisions, where the inputs are far messier. You are not placing a bet with a fixed edge. You are deciding how much of your net worth to allocate to a house, a business, a partner, or a career pivot with information that will change after you commit. I ran into this when someone asked me about putting 40 percent of their liquid assets into a friend's startup. Standard Kelly would have said zero, because the expected value calculation was too noisy. But ignoring it entirely meant they walked away from a legitimate asymmetric opportunity. The workaround was to treat the Kelly result as a ceiling, not a floor, then apply a 1/4 Kelly filter with a hard stop at 15 percent of total portfolio value. That got the exposure down to something survivable while keeping the option alive. I used a simple spreadsheet with three scenarios—base case, downside, and best case—weighted by my best guess at probability. The spreadsheet output suggested 8 percent. I rounded down to 6 percent because I knew startup returns have fat left tails that Kelly underweights.

The Core Mechanics of Kelly Love And Money

At its foundation, Kelly sizing comes from this equation: f* = (bp - q) / b, where f* is the fraction of your bankroll to wager, b is the odds received (net payout per unit wagered), p is the probability of winning, and q is 1 minus p. This is not a suggestion. It is the mathematical answer to maximizing the geometric growth rate of your wealth over repeated bets. Ed Thorpe and Claude Shannon proved this in the 1960s. John Kelly originally derived it for telephone line capacity at Bell Labs before anyone connected it to gambling. The reason people fail with Kelly in real life is not the formula. It is the inputs. In a blackjack game, you can count cards and get a reasonable estimate of p. In your life, p is something you estimate from gut feel, market research, and past experience that may not repeat. A house purchase is not a binary bet. A romantic partnership is not a binary bet. A side business with three years of runway is closer to a binary bet, but even then, the payout distribution is highly skewed. Kelly assumes you know the payout distribution. You do not. Here is the part most people skip. The Kelly fraction assumes you can rebalance after every round. In personal finance, you cannot. When you buy a house, that capital is locked. When you invest in a company, your stake may not be liquid for five years. This illiquidity penalty means you should scale down the Kelly fraction significantly. A common adjustment used by practitioners is to apply a fractional Kelly multiplier between 0.25 and 0.5. This reduces volatility without sacrificing most of the long-term growth advantage. Full Kelly in a real-world setting with concentrated, illiquid bets will blow up your portfolio before the math has time to work.

Applying It to Major Life Decisions

Let me walk through a realistic example. Say you have $500,000 in investable assets and you are considering a $150,000 investment in a small business venture. You estimate a 30 percent chance the business succeeds and returns 4x your money, a 40 percent chance it breaks even, and a 30 percent chance it goes to zero. The expected value is positive: 0.30 times 4 minus 0.70 equals 0.50, or a 50 percent expected return. Plugging into Kelly with b equal to 3 (you profit 3 units for every 1 unit wagered) and p equal to 0.30, the formula gives you roughly 10 percent of your bankroll. That is $50,000. The full $150,000 ask exceeds what Kelly recommends by a factor of three. Now here is the uncomfortable part nobody tells you. Kelly does not account for correlation. If you already own a home, have a mortgage, and work in a thin industry, this business may actually be correlated with your human capital. A recession hits your job and the business simultaneously. Kelly treats each bet as independent. It is not. I learned this the hard way when a client followed Kelly sizing precisely and then lost 60 percent of his portfolio in eighteen months because three of his "independent" bets failed during the same downturn. The fix was introducing a correlation adjustment factor that reduced each position size by 30 to 40 percent when the underlying risk drivers overlapped. It was boring. It worked. When you apply this to relationships, the framing shifts but the math does not disappear. Investing time, emotional energy, and shared assets into a partnership is a form of capital allocation. The Kelly principle reminds you to size your commitment relative to the quality of your information and the reversibility of the decision. A cohabitation decision with someone you have known for six months is a high-variance bet with low information. Kelly would argue for minimal commitment. A marriage decision after five years of shared finances, conflict resolution history, and aligned goals carries more informational density. The Kelly fraction for that bet can be higher. This does not mean love is quantifiable. It means the discipline of asking yourself whether your input matches the certainty level is useful.

Common Pitfalls and What to Do Instead

Pitfall one: confusing arithmetic mean with geometric mean. Kelly maximizes geometric growth, which is always lower than arithmetic growth when variance is present. People see a strategy with a 20 percent average annual return and assume they will get 20 percent compounding. They will not. The actual compound rate depends on the variance of returns. High-variance strategies look great on paper and devastate investors in practice. The workaround is to run a Monte Carlo simulation with your estimated return distribution and observe the median outcome, not the average. You will almost always prefer a lower-mean, lower-variance strategy. Pitfall two: overconfidence in probability estimates. Human beings are terrible at assigning probabilities to unique events. You think you have a 70 percent chance of success on a project. The track record says you are wrong 40 percent of the time when you feel that confident. The workaround is calibration training. Keep a simple log of your predictions and outcomes. After fifty entries, compare your stated probabilities to actual hit rates. Most people discover they are miscalibrated by 15 to 25 percentage points. Adjust your estimates toward the mean after you see the data. Pitfall three: ignoring drawdown tolerance. Kelly will suggest bet sizes that create drawdowns you cannot emotionally withstand. Even if the math works over ten years, you may panic-sell at the bottom of year three. I have seen this repeatedly. The practical solution is to introduce a maximum drawdown constraint. If a Kelly-optimal position would expose you to a 35 percent portfolio decline, cut the position size until the modeled maximum drawdown falls below your personal threshold. Most people find 20 to 25 percent comfortable. Anything above that requires psychiatric-level discipline most people do not possess.

Pitfall four: applying Kelly to non-repeating decisions. Kelly assumes repetition. A single home purchase is not repeated. A single business venture is not repeated. Using Kelly literally for one-off decisions gives you a false sense of precision. The workaround is to treat Kelly as a rough directional guide rather than a answer. Ask whether the Kelly fraction is telling you to go all in, stay out, or bet modestly. The direction matters more than the decimal point.

A Realistic Workflow for Using Kelly Love And Money

Step one: define the decision as a bet with identifiable payoffs. Write down the possible outcomes and their monetary values. Be specific. "The business might do well" is not an input. "There is a 25 percent chance of a 5x return, a 50 percent chance of a 1.5x return, and a 25 percent chance of total loss" is an input. Step two: assign probabilities to each outcome. Use historical data when available. Use market analogs when you lack direct data. When you truly have no information, admit it and assign a uniform distribution or step back from the bet entirely. Step three: calculate the Kelly fraction. Use the formula or a calculator. Apply a fractional Kelly multiplier of 0.25 to 0.5 depending on your confidence in the inputs and the liquidity of the asset.

Step four: stress test with correlation and drawdown constraints. Run worst-case scenarios. Check whether the position size respects your maximum drawdown tolerance. Adjust downward if needed. Step five: implement with a hard stop. Define the exit condition before you enter. If the business misses two consecutive milestones, you sell. If the relationship passes three specific compatibility markers within twelve months, you deepen commitment. Without a precommitment rule, you will rationalize staying in a losing bet longer than Kelly would ever recommend. The Kelly Love And Money framework is not a crystal ball. It is a discipline for preventing the most common mistake people make with their resources: betting too large on low-information opportunities and too small on high-information ones. The math is simple. The application is hard. The people who get it right are the ones who respect both the formula and its limitations.