Among the most well-known bet optimization methods is the Kelly Criterion. Named after its inventor, John L. Kelly of AT&T Bell Laboratories, the Kelly Criterion is a formula designed to boost betting profits by determining the ideal percentage of capital to be staked on each wagering event.

The formula is as follows:

WINNING RATE – LOSING RATE / AVG. WINNING ODDS
(Note: Remember your order of operations — division before subtraction.)

Hence, if one typically wins 30 percent of her bets at average winning odds of 5-2, the amount wagered to optimize profits would be approximately two percent of capital (0.3 – 0.7 ÷ 2.5 = 2 percent).

Because there is such risk of tapping out if one overestimates his true advantage, most bettors use a fraction of Kelly to determine their stake.

Now, it should be clear from this explanation that the Kelly Criterion is used only to exploit an edge, not to create one. In other words, if the Kelly Advantage derived from the formula above is negative, you don’t have an edge; you’re better off trusting in Fortuna, the Roman god of fortune and chance, than Kelly.

For example, let’s say that through record-keeping, you discover that you cash 30 percent of your win bets (like above), but at average winning odds of 2-1. This yields a -5 percent Kelly Advantage… which is no advantage at all.

Please read this carefully: You cannot, via any wagering method, scheme or spiritual revelation, turn a negative return on investment (ROI) into a positive one. The Kelly Criterion only works if one has an advantage over the game they are playing. (You can turn a negative ROI into a positive one — at least in games of skill like horse racing — through greater education and better information.)

Of course, this begs the question: Why use the Kelly Criterion at all? If one already has an edge, what difference does it make whether he follows the advice of John Kelly or Kelly Clarkson? The answer is simple: It gives one an idea as to how much she should bet on each race to maximize profits (again, assuming one has an inkling as to what her betting statistics are — and they show a positive return).

For instance, let’s say you plan on risking up to $1,000 on the card at Belmont Park today. The first race is for 3-year-old state-bred maidens and your records show that you typically hit 20 percent winners at average odds of 3-1 in such affairs. Plugging those numbers into the Kelly calculator, we find that your “advantage” is -6.67 percent. This means you have no edge in the race and should pass (or bet very lightly if you’re just playing for fun).

The second race is a $40K claiming sprint for older horses — you love this race and have a “best bet” named Ima Winner. Traditionally, your best bets have hit at a 40 percent rate with average winning odds of 2-1 — a whopping 10 percent Kelly edge. Hence you wager $100 (10 percent of your $1,000) bankroll. Ima Winner lives up to its name and pays $6.40. You collect $220 in profit and your bankroll now stands at $1,220.

In the third event on the Belmont program, you find an angle play that has a 15 percent strike rate and 8-1 average winning odds. This equates to a 4.38 percent Kelly bet, or $53 of your current bankroll ($1,220 x 4.38% = $53).

Unfortunately, the angle play finishes last and your bankroll now stands at $1,167.

You continue sizing your bets in this manner for the rest of the card, adjusting your bankroll as you go.

Again, it needs to be stressed that the Kelly Criterion is not a handicapping tool, it just helps one make better betting decisions.

Still, it’s a lot of fun to use and can really help those who too often find themselves on the losing side of the ledger after a great day of handicapping.