The Kelly Criterion for Value Bettors: Staking When You Have an Edge

Finding value is half the battle. The other half is staking correctly on that value. The Kelly Criterion is a mathematical formula that tells you how much of your bankroll to bet on any given edge. Get it right and your bankroll compounds at the fastest possible rate. Get it wrong and you risk catastrophic losses. This is why understanding Kelly is non-negotiable for serious value bettors.

How the Kelly Criterion Works

The Kelly Criterion was developed by John Kelly at Bell Labs in 1956 to optimize information transmission in noisy communication channels. It was later adopted by gamblers and investors. The formula is elegant:

f = (bp - q) / b

Where:

• f is the fraction of your bankroll to bet
• b is the decimal odds minus one (the net odds)
• p is your probability of winning
• q is your probability of losing (1 - p)

Let's work through a practical example. You've identified a value bet where you think the true probability of a team winning is 55%. The bookmaker is offering 2.10 odds (implied 47.6% probability).

First, calculate the edge. The fair value odds for 55% probability are 1/0.55 = 1.82. The bookmaker is offering 2.10, which is better. You have an edge.

Now apply Kelly:

• b = 2.10 - 1 = 1.10 (net odds)
• p = 0.55 (your probability)
• q = 0.45 (1 - 0.55)

f = (1.10 * 0.55 - 0.45) / 1.10 f = (0.605 - 0.45) / 1.10 f = 0.155 / 1.10 f = 0.141 or 14.1%

According to full Kelly, you should bet 14.1% of your bankroll on this bet. If your bankroll is £1,000, you'd stake £141.

Why Kelly Works for Bankroll Growth

Kelly is optimal because it maximises the long-term growth rate of your bankroll. No other staking strategy can achieve faster growth given the same edge and odds. This is proven mathematically.

Here's why. Imagine you have an edge where you win 55% of the time at 2.10 odds. Over 100 bets:

• You win 55 bets, each returning 2.10 times your stake.
• You lose 45 bets, losing your stake.

Net result: 55 * 2.10 - 45 = 115.5 - 45 = 70.5 for every 100 units wagered. That's a 70.5% return on total action.

Kelly tells you the optimal proportion of your bankroll to risk on each bet to achieve the highest compound growth rate. Risk too little and you're not taking full advantage of your edge. Risk too much and you risk ruin. Kelly finds the sweet spot.

Worked Examples: Kelly Across Different Scenarios

Scenario 1: Strong Edge, High Odds

• Your probability: 60%
• Bookmaker odds: 3.00
• Fair value odds: 1/0.60 = 1.67

b = 3.00 - 1 = 2.00 p = 0.60 q = 0.40

f = (2.00 * 0.60 - 0.40) / 2.00 f = (1.20 - 0.40) / 2.00 f = 0.80 / 2.00 f = 0.40 or 40%

Kelly says bet 40% of your bankroll. This is a strong edge and high odds combine to warrant aggressive sizing.

Scenario 2: Weak Edge, Low Odds

• Your probability: 52%
• Bookmaker odds: 1.60
• Fair value odds: 1/0.52 = 1.92

b = 1.60 - 1 = 0.60 p = 0.52 q = 0.48

f = (0.60 * 0.52 - 0.48) / 0.60 f = (0.312 - 0.48) / 0.60 f = -0.168 / 0.60 f = -0.28 or negative

Kelly says don't bet. The edge is too small relative to the odds. A 52% probability at 1.60 odds is not value. You need the odds to reflect better than your probability estimate, and by a meaningful margin. This is Kelly's built-in safety mechanism.

Scenario 3: Moderate Edge, Fair Odds

• Your probability: 55%
• Bookmaker odds: 2.00

b = 2.00 - 1 = 1.00 p = 0.55 q = 0.45

f = (1.00 * 0.55 - 0.45) / 1.00 f = (0.55 - 0.45) / 1.00 f = 0.10 / 1.00 f = 0.10 or 10%

Kelly says bet 10% of your bankroll. A moderate edge warrants moderate sizing.

Why Full Kelly Is Too Aggressive in Practice

The Kelly formula assumes perfect knowledge of your edge. In reality, you estimate your edge based on analysis. Your estimate might be wrong. You might think a team has a 55% chance but it's actually 52%. If you're betting full Kelly based on an overestimated edge and your true edge is lower, you'll lose money faster.

This is why professional bettors use fractional Kelly, typically 25% Kelly, 50% Kelly, or even lower.

25% Kelly: You bet one quarter of what full Kelly recommends. If full Kelly says 40%, you bet 10%. This cushion protects you against overestimating your edge. You still achieve long-term bankroll growth, just slower. But the reduction in volatility is worth it.

50% Kelly: You bet half of what full Kelly recommends. For a 40% full Kelly edge, you'd bet 20%. This is popular among professional bettors who are confident in their edge estimation but want to reduce volatility.

The trade-off is simple: fractional Kelly gives you slower bankroll growth in exchange for lower risk of ruin and better sleep at night. Most professional bettors prefer this.

How to Implement Kelly in Your Betting

Step 1: Identify your edge. Use all available information: xG data, team form, underlying metrics, and your own analysis. Estimate the true probability of the outcome.

Step 2: Calculate fair value odds. Convert your probability to odds: odds = 1 / probability. If you think a team has 60% chance to win, fair value is 1.67.

Step 3: Check bookmaker odds. Is the bookmaker offering odds better than fair value? If yes, you have a positive edge. Calculate the edge size.

Step 4: Apply Kelly formula. Use the formula to calculate the percentage of your bankroll you should bet. Be conservative: use 25-50% Kelly, not full Kelly.

Step 5: Calculate stake in pounds. If your bankroll is £2,000 and Kelly recommends 15%, you stake £300.

Step 6: Place the bet. Track the result and adjust your edge estimation based on outcomes. Over time, your edge estimation should improve.

Step 7: Rebalance as your bankroll changes. As your bankroll grows or shrinks, your stake amounts should adjust proportionally.

The Mechanics of Fractional Kelly

Using fractional Kelly is simple arithmetic. If full Kelly recommends 20% and you want to use 50% Kelly, multiply: 20% * 0.50 = 10%. You stake 10% instead of 20%.

The long-term growth rate is proportional to the Kelly percentage you use. At 100% Kelly (full Kelly), you maximise growth. At 50% Kelly, you get roughly 50% of the growth rate but significantly lower volatility. At 25% Kelly, growth is slower but volatility is even lower.

Most serious value bettors settle on 25-50% Kelly as a practical compromise. Growth is still fast enough to be meaningful, but risk is manageable.

The Dangers of Ignoring Kelly or Misapplying It

Betting the same amount on all bets: This ignores your edge size. A bet where you have 15% edge should be larger than a bet where you have 2% edge. Equal staking leaves money on the table.

Betting full Kelly on estimated edges: If your edge estimation is off by just a few percentage points, full Kelly can lead to catastrophic losses. This is the most common mistake. Always use fractional Kelly as insurance.

Using a fixed Kelly percentage for all leagues and markets: Your edge varies. Better edges deserve bigger bets. A fixed Kelly percentage is more practical but less optimal than recalculating for each individual bet.

Ignoring Kelly once you're up: As your bankroll grows, many bettors stop using Kelly discipline and start betting randomly. This wastes the growth you've earned. Keep using Kelly.

In Summary

• The Kelly Criterion is a mathematical formula for optimal bet sizing based on your edge: f = (bp - q) / b
• The formula maximises long-term bankroll growth by staking more on larger edges and less on smaller edges
• Full Kelly assumes perfect edge knowledge, which is unrealistic; overestimating your edge leads to catastrophic losses
• Fractional Kelly (25%-50% of full Kelly) is safer in practice and protects against overestimating your edge
• The process: estimate true probability, calculate fair value odds, check if bookmaker offers better odds, apply Kelly formula
• Kelly staking combined with proper edge identification produces exponential bankroll growth
• Start with 10-25% Kelly if new to value betting; gradually increase to 25-50% Kelly as your track record improves
• Kelly is only optimal if your probability estimates are accurate; if your estimates are off, fractional Kelly provides protection
• Without Kelly discipline, you're betting randomly on the same stakes regardless of edge quality

FAQ

Q: How much Kelly should I use if I'm new to value betting? A: Start with 10-25% Kelly. This is conservative but protects you while you develop skill in edge estimation. As your confidence and track record improve, gradually increase to 25-50% Kelly.

Q: What if Kelly tells me not to bet? A: That's correct. If the formula produces a negative number, there's no edge. Don't bet. This is one of Kelly's key strengths: it filters out bets that look like value but mathematically aren't.

Q: Can I use Kelly across different sports or markets? A: Yes. Kelly works whenever you have estimated probability and bookmaker odds. The principle is the same for football, tennis, horse racing, or any other market. Your edge estimation ability is what varies.

Q: Does Kelly work for live/in-play betting? A: Kelly works mathematically, but implementing it in-play is harder because odds change constantly. For fast-paced in-play markets, many bettors use fixed percentage stakes instead.

Q: How do I calculate Kelly if I'm betting multiple correlated events? A: This is advanced territory. For your first years of betting, assume independence and apply Kelly to individual bets. Once you have a large track record, you can model correlations and refine further.

Q: What bankroll should I start with for Kelly betting? A: Large enough that your smallest Kelly bet represents a realistic unit for your bookmaker. If Kelly recommends 1% of a £500 bankroll (£5) and your bookmaker has a £5 minimum stake, that works. If minimum stakes are £10, start with a larger bankroll.