Choosing an allocation strategy
You have done the hard work: found a system, identified edges, and built a portfolio. Now comes a question that will quietly determine a huge portion of your results: how much should you invest in each position? The two most common answers are flat trading and the Kelly Criterion, and they represent fundamentally different philosophies about risk, growth, and simplicity.
Neither is universally superior. The right choice depends on your experience level, your confidence in your probability estimates, and your psychological tolerance for drawdowns. Let us break down both approaches with real numbers so you can make an informed decision.
Flat trading: simplicity and safety
Flat trading means staking the same fixed amount on every position, regardless of your perceived confidence in the outcome. If your unit is $20, every position is $20 — whether you see it as a strong play or a marginal edge.
The primary advantage is simplicity. There are no calculations to make, no subjective confidence ratings to assign, and no risk of catastrophically overbetting on a play where your edge estimate is wrong. Flat trading also produces the smoothest equity curve: your portfolio grows (or shrinks) in small, predictable increments.
• Zero complexity — no formulas, no subjective inputs
• Lowest variance of any allocation method
• Psychologically easy to follow during losing streaks
• Immune to errors in probability estimation
Cons of flat trading:
• Does not capitalize on higher-edge opportunities
• Slower portfolio growth compared to variable allocation
• A $20 invest in a 10% edge play and a 2% edge play is not capital-efficient
Kelly Criterion: mathematically optimal growth
The Kelly Criterion is a formula developed by John Kelly at Bell Labs in 1956. It calculates the optimal fraction of your portfolio to allocate based on your estimated edge and the returns offered. The goal is to maximize the long-term growth rate of your portfolio.
f* = (bp − q) / b
Where:
f* = fraction of portfolio to allocate
b = decimal returns minus 1 (the net payout per dollar staked)
p = your estimated probability of winning
q = your estimated probability of losing (1 − p)
Example: Returns of 2.20, estimated 50% win probability.
b = 1.20, p = 0.50, q = 0.50
f* = (1.20 × 0.50 − 0.50) / 1.20 = 0.10 / 1.20 = 8.3% of portfolio
Kelly is mathematically proven to maximize the geometric growth rate of your portfolio over time, assuming your probability estimates are accurate. It naturally positions more on high-edge situations and less (or nothing) on low-edge or no-edge situations.
• Maximizes long-term growth rate (mathematically proven)
• Automatically sizes positions proportional to your edge
• Naturally avoids overbetting when edge is small
• Positions zero when there is no edge
Cons of Kelly:
• Requires accurate probability estimates (the hardest part of trading)
• Full Kelly produces extremely high variance
• Overestimating your edge by even a small amount leads to systematic overbetting
• Psychologically demanding during drawdowns
The problem with full Kelly
In theory, full Kelly is optimal. In practice, it is dangerously aggressive for almost everyone. The reason is simple: the formula assumes your probability estimates are perfectly calibrated. They never are.
If you estimate a 55% probability but the true probability is 52%, full Kelly will have you trading significantly more than the actual optimal amount. Over time, this systematic overbet erodes your portfolio instead of growing it. The formula is extremely sensitive to input accuracy.
Additionally, full Kelly produces gut-wrenching drawdowns. Simulations show that a full Kelly investor will typically experience a 50% drawdown from peak at some point, even with a genuine edge. Most humans cannot sit through that without abandoning their strategy.
Fractional Kelly: the practical middle ground
Most experienced investors who use Kelly employ a fractional version — typically quarter Kelly (25%) or half Kelly (50%). This means you calculate the full Kelly allocation and then position only a fraction of it.
- Quarter Kelly (25%): Very conservative. Growth is slower but drawdowns are small and manageable. Best for investors who are uncertain about the accuracy of their probability estimates.
- Half Kelly (50%): The most common choice among professional investors. It sacrifices only about 25% of the growth rate compared to full Kelly while cutting drawdowns roughly in half. A good balance of growth and safety.
- Three-quarter Kelly (75%): More aggressive. Only appropriate if you have strong confidence in your probability estimates, typically backed by a proven model with a long track record.
For a more detailed look at the Kelly Criterion, unit sizing, and other portfolio concepts, see our portfolio management guide.
Which should you choose?
Here is a practical decision framework:
- Choose flat trading if: You are relatively new to sports investing, you do not have a statistical model generating probability estimates, you prefer simplicity, or you find drawdowns stressful.
- Choose fractional Kelly if: You have a model or methodology that produces probability estimates you trust, you understand that drawdowns will be larger, and you want to maximize growth over the long run.
- Never use full Kelly unless you are running a quantitative operation with rigorously backtested probability models and you have deep experience managing variance.