The Kelly Criterion Explained (Without the Math Headache)

The sizing formula used by hedge funds, pro poker players, and the sharpest American bettors — broken down in plain English with real-world NFL and NBA examples.

In Post #4, Bankroll Management 101, we laid out the unit system — the simple, reliable framework that protects American bettors from going broke. But the unit system has a built-in limitation: it treats every bet the same regardless of how big your edge actually is. That's where the Kelly Criterion comes in.

Originally developed in 1956 by John Kelly, a researcher at Bell Labs, the Kelly Criterion is a mathematical formula that tells you the optimal percentage of your bankroll to wager on any given bet — based on the size of your edge. It's been used by legendary investors like Warren Buffett's mentor Ed Thorp, by professional poker players, and by the sharpest sports bettors in the world.

The good news? You don't need a math degree to use it. By the end of this post, you'll understand exactly how Kelly works, when to use it, when to scale it down, and why most American bettors should probably never use the full version.

01 What Problem Does Kelly Actually Solve?

Imagine you've found a bet you genuinely believe gives you a real edge. Maybe FanDuel has the Lakers at +150 in a game where you've calculated they should be closer to +110. You know the bet has value — but how much should you actually wager?

The standard unit system says: bet 2% of your bankroll. But that ignores something critical — the edge itself. A massive edge deserves a bigger bet than a tiny one. Betting the same amount on a 0.5% edge and a 10% edge is mathematically inefficient. You're leaving money on the table on the big edges and risking too much on the small ones.

Kelly fixes this. It scales your bet size proportionally to how much edge you actually have — bigger bets on bigger edges, smaller bets on smaller ones. Done correctly, it's the mathematically optimal way to grow a bankroll over time.

02 The Kelly Formula (In Plain English)

Here's the formula. Don't panic — it looks more intimidating than it actually is.

Translated into bettor English: multiply your win probability by the potential payout, subtract your loss probability, then divide by the payout. The result is the percentage of your bankroll to wager.

It sounds abstract until you plug in real numbers. Let's do that now.

03 A Real-World Kelly Example (NFL Moneyline)

Imagine it's Sunday morning. You're looking at a Bengals vs. Browns game. DraftKings has the Bengals at +150 on the moneyline. After deep research — injuries, weather, line movement, recent form — you've concluded the Bengals actually have about a 50% chance to win.

Quick check: at +150, the implied probability is 40%. Your model says 50%. That's a real edge — about 10 percentage points. So how much should you bet?

Plug it into Kelly:

• b = +150 in decimal is 2.50, minus 1 = 1.50
• p = your win probability = 0.50
• q = your loss probability = 0.50

Now run the math:

(1.50 × 0.50) − 0.50 = 0.75 − 0.50 = 0.25

0.25 ÷ 1.50 = 0.1667

That's 16.67% of your bankroll. On a $2,000 bankroll, full Kelly says to bet $333.33 on that Bengals moneyline.

04 Why Full Kelly Is Brutal in Practice

Full Kelly is mathematically optimal — but only if two conditions are true:

1. You know your true edge with perfect accuracy.
2. You're emotionally capable of withstanding massive bankroll swings.

Both of those conditions are nearly impossible to meet in real-world sports betting. Here's why:

Edge Estimation Is Hard

You can't actually know if the Bengals have a 50% chance to win. You're estimating. If your real edge is 5% instead of 10%, full Kelly will tell you to overbet by a wide margin. That overbet compounds across many bets, and you end up risking far more than you should.

The Variance Is Insane

Full Kelly bettors can experience 50% bankroll drawdowns as a normal occurrence — not a worst-case scenario. Studies show that betting at full Kelly has roughly a 1-in-3 chance of cutting your bankroll in half before doubling it. Most bettors mentally cannot handle that level of swing without going on tilt.

05 Fractional Kelly — The Sharp Bettor's Approach

This is where the rubber meets the road. Almost no professional uses full Kelly. Instead, they use fractional Kelly — typically quarter Kelly (0.25x) or half Kelly (0.5x).

The idea: take whatever full Kelly tells you to bet, then bet a fraction of that. So in the Bengals example above:

• Full Kelly = 16.67% of bankroll = $333.33 on a $2,000 roll
• Half Kelly = 8.33% of bankroll = $166.67
• Quarter Kelly = 4.17% of bankroll = $83.33
• Eighth Kelly = 2.08% of bankroll = $41.67

Notice how quarter Kelly produces a bet size right around the standard 2–4% range we covered in Post #4. That's not a coincidence — that's the sweet spot where the math of Kelly and the practical wisdom of the unit system converge. Most sharp American bettors live in the quarter Kelly to half Kelly range.

06 When Kelly Tells You to Bet Zero (or Negative)

Here's the most useful feature of the Kelly Criterion that almost no one talks about: it tells you when not to bet.

If the formula produces a result of zero or a negative number, that means you have no edge — and Kelly is telling you to pass entirely. This is huge. Most American bettors place way too many bets, betting on games where they don't actually have an edge, just because the games are exciting or they want action.

Example: BetMGM has the Cowboys at -110 against the Eagles. Your model says the Cowboys have exactly a 52.4% chance to win — which is the break-even threshold at -110. Plug it into Kelly:

• b = -110 in decimal is 1.91, minus 1 = 0.91
• p = 0.524
• q = 0.476

(0.91 × 0.524) − 0.476 = 0.477 − 0.476 = 0.001

0.001 ÷ 0.91 = 0.0011

That's a Kelly recommendation of 0.11% of your bankroll — essentially zero. The formula is correctly telling you: yes, you might have a tiny edge here, but it's so small that betting on it isn't worth the variance. Pass.

This discipline alone — the willingness to skip marginal bets — separates winning American bettors from the rest of the public.

07 The Kelly Cheat Sheet for American Odds

You don't need to crunch numbers every time. Here's a quick reference for common scenarios — assumes the odds you're getting versus your estimated true probability:

Pin this table somewhere. For most American bettors using quarter Kelly, your bet sizes should almost always fall between 1% and 5% of your bankroll. If you're getting Kelly numbers higher than that, double-check your edge estimate — you're probably being overconfident.

08 Kelly's Biggest Trap: Overestimating Your Edge

The Kelly Criterion is only as accurate as your win probability estimates. And here's the uncomfortable truth: most American bettors overestimate their edge by a wide margin.

You think you have a 55% chance to hit a bet. Reality might be 51%. That's a 4-point gap — and at full Kelly, that gap turns a "smart" bet into a money-losing one. Combine that with the variance involved in sports betting, and overestimating your edge is the fastest way to destroy a bankroll while doing math that looks completely rigorous.

09 Should You Actually Use Kelly?

Here's the honest answer for most American bettors:

Don't Use Kelly If…

• You're new to sports betting (less than 500 tracked bets).
• You don't have a documented method for estimating win probabilities beyond gut feel.
• You can't emotionally handle a 25–50% bankroll swing.
• You bet on parlays, same-game parlays, or props as your main strategy. Kelly doesn't work well for these (we'll cover why in a future post).

Do Consider Kelly If…

• You've been tracking bets for at least a full season and can prove you're winning.
• You have a systematic approach to building win probability estimates — power ratings, models, or strict criteria.
• You're betting primarily straight bets (moneylines, spreads, totals) at major U.S. sportsbooks.
• You'll commit to fractional Kelly — quarter or half — not full Kelly.

If you're not in the second group yet, stick with the unit system from Post #4. It's not as mathematically optimal as Kelly, but it's far more forgiving of estimation errors — and those errors are the real killer.

10 A Hybrid Approach for Most Bettors

Here's a practical compromise that captures most of Kelly's benefit without the variance: use Kelly's logic, but cap your maximum bet size.

For example, on a $2,000 bankroll:

1. Calculate your edge for each potential bet.
2. Use quarter Kelly to determine your bet size.
3. Cap any single bet at 3% of your bankroll ($60 in this example), no matter what Kelly says.
4. Skip any bet where Kelly says less than 1% — that's not enough edge to be worth the variance.

This hybrid approach gives you the proportional sizing benefit of Kelly (bigger bets on bigger edges) without the catastrophic risk of trusting overconfident edge estimates. It's what a lot of professional American bettors actually do in practice — even if they describe themselves as "Kelly bettors."

Final Thoughts — Math Is a Tool, Not a Crutch

The Kelly Criterion is one of the most powerful concepts in sports betting math. Used correctly, it can meaningfully accelerate the growth of a winning bankroll. Used incorrectly — or used by a bettor who isn't actually winning — it can blow up an account faster than you'd believe.

The biggest mistake American bettors make with Kelly isn't the math. It's assuming that doing the math means they actually have an edge. The formula doesn't create an edge. It just helps you size correctly if you already have one.

Start with the unit system. Track your bets for hundreds of wagers. Prove to yourself that you're actually winning. Then, and only then, layer in Kelly's logic to sharpen your sizing. That's the path the sharpest bettors actually walk — and it's the one most likely to keep you profitable for the long haul.