Kelly Criterion

Determining how much you stake on a wager is a crucial consideration for successful punters. The Kelly Criterion is a staking method well known across wagering and investment professionals which should be known and considered by all Betfair punters.

Proportional Staking 

This involves increasing your stakes proportionately to how “wrong” you think the odds are. In a very general terms, if a horse is $7.00, when you think it should be $5.00, you might have $10 on it. But if the same horse were $12.00 then you should increase your bet size.

For many punters, this can feel counter-intuitive. Betting at bigger odds means that you don’t have to put as much down to get an attractive return, so why would you risk more? And anyway, don’t the bigger odds mean that you stand less chance of winning?

Well, in some respects, yes. But the point is that, as a punter, you have to strike when things are most in your favour. If you were offered $2.10 on a single coin toss, you might be tempted to have a bet. If you were offered $3.00, wouldn’t you bet more?

The question becomes, then: how much more?

Minimum & Maximum Stake Limits 

The issue with betting big when the odds are most in your favour means that it can only take a run of a few losing bets – which will happen no matter how much the odds are in your favour: heads sometimes comes up ten times in a row – to completely wipe out your betting bank.

To combat this, most successful punters operate some kind of system where they have a minimum and maximum stake limit – say $10 to $40 – and they select the size of bet depending on how much value they think they are getting. In these cases, the upper limit usually represents a chunk of their betting bank in the range of 2% to 5%. This means that they can endure a losing streak without having their funds wiped out. How big a percentage you are prepared to risk on your biggest bets will depend largely on your attitude to risk, but once that upper limit is set, stick to it.

Kelly Criterion Staking Strategy 

One way of becoming ever more sophisticated in this bet-more-when-the-odds-are-in-your-favour approach is to adopt a strategy known as the Kelly Criterion. This was developed by a mathematician called John Kelly in the 1950s. Under analysis it has shown to be an effective way of maximising profit when operated in conjunction with a successful selection strategy.

The Kelly formula is: (BP-Q) / B

Where:

B is the odds you are getting.  With decimal odds, you need to subtract 1.

P is the likelihood of the bet winning.

Q is the probability of losing, or 1 – P.

If you thought a horse should be 1.60 (which translates to 62.5% in probabilistic terms, (1/1.60)), but it was available to back at 1.70, then the sum would look like this:

((.70 x 0.625)-0.375) / .70 = 8.93%

Meaning that you should place 8.93% of your betting bank on the outcome.

However, the example above throws up a couple of immediate concerns: the size of bets that can be recommended and the need for an accurate prediction of probability.

Size of Your Bets 

 First, and most pressing, is that when the odds are really in your favour, the Kelly Criterion calls for you to bet significant portions of your betting bank, which is difficult to swallow for even the most aggressive punters, no matter what the mathematical proofs may tell them. To guard against the volatility that the Kelly Criterion can bring to their betting bank, most successful punters using the approach will either set a maximum upper limit for their bets, or always bet a fixed fraction of what the formula suggests.

Part of the reason for this is that the Kelly Criterion relies on a mathematically certain prediction of chance, and whilst this can be provided for a coin toss, accurately assessing the chance of a sporting outcome is usually fraught with a greater degree of uncertainty.

Accurate Prediction Models 

Whether adopting a full or partial Kelly approach, or a simpler bet-more-when-the-odds-are-in-your-favour approach, the fact remains that all staking strategies are useless unless you have reliable ways of working out what odds a sporting outcome should be.

There are lots of ways to do this – with the use of ratings systems being the most popular – but this should be the challenge that punters take on before they worry about staking strategies, no matter what the internet forums might suggest.

Related Articles