Staking – Level Staking

Problem One

First, they require punters to be able to accurately assess the probability of something happening and compare this number to the odds they are getting. If you are tossing a coin, this is easy, because you are dealing with the mathematical certainty that heads will come up, in the long run, 50% of the time. If you are then offered odds of $3.50, which imply heads will only come up 29% of the time, it is relatively simple to work out how “wrong” the odds are and bet accordingly.

It is no surprise that the Kelly Criterion became popularised in betting circles by blackjack-playing card-counters like Edward Thorpe. The basis of his blackjack technique – outlined in his famous book, Beat the Dealer – was to determine the extent to which the deck of cards was stacked in his favour and bet accordingly. When you are dealing with reliable and specific numbers – like the fact that there are 52 cards in a deck – then reliably assessing probability is straightforward.

Most sports betting, however, is not like this: the nature of it means that you cannot be mathematically certain about anything. For example, you may think that odds of $1.76 on Melbourne beating St Kilda represent value, but it’s hard to come up with the precise probability of them winning, which is what systems like the Kelly Criterion require. In Australian Rules, as in all sports, there are just too many unquantifiable variables to have that kind of certainty.

Many punters try to get around this by using rating systems and statistical models that provide greater certainty about the predictions they are making. And it’s important to note that, if you can do this – even if you are not completely accurate – the Kelly Criterion and other proportional staking systems are still applicable to sports betting, which is why many professional punters base their staking strategies around some form of them.

Problem Two

The second disadvantage is that betting proportionately to how wrong the odds are can be an emotional adventure. Using the “pure” Kelly Criterion can often mean investing significant proportions of your betting bank in one bet. If a few of these bets come together – and lose (which will happen from time to time) – then you can find yourself doubting everything. It’s all very well telling yourself that the Kelly

Criterion is mathematically proven to maximize profits in the long run, but because you’re not dealing with the mathematically certain probabilities involved in a coin-toss, you’ll start mistrusting the bets you are making. It takes an inhuman amount of psychological detachment to remain unemotional through such a losing streak, which is why most punters using the Kelly Criterion will modify it in some way by capping the maximum bet that they will place.

What Does This Mean?

These disadvantages mean that – when it comes to sports betting – many punters choose a more straightforward strategy: level staking. For example, they decide that they will bet $20 on every bet, every time.

This doesn’t feel especially sophisticated – after all, if you’re not using a spreadsheet to work out your stake, how good can it be! And yet, in various simulations, providing you have a way of identifying value bets, level-staking provides consistent returns that outperform many (so-called) strategies like Labouchere, Martingale and D’Alembert (which usually end in complete ruin).  And although it might not provide the stellar returns of the Kelly Criterion, it does avoid the associated volatility.

So, if you want a staking system that is mathematically sound, and you are using a selection criteria which means you might be able to identify a value bet without necessarily being able to precisely assess the exact probabilities involved, then level staking is the way to go.

That’s not to say that it doesn’t have some disadvantages, though. Aside from the fact that it does not fully capitalise on occasions when the odds you are getting are most wrong, like the Kelly Criterion and other proportional systems do, it also doesn’t take account of either the size of your betting bank or the type of bet you are making.

If you suffer a run of losing bets, the level stake you are using would represent a much bigger percentage of your betting bank than it initially did, which therefore represents a riskier betting proportion: a $50 bet is conservative when you have a $4,000 bank; it’s downright buccaneering when you are down to your last $100.

Also, most punters make a wide variety of different types of bets: from the long-odds speculative punt to the near-certain odds-on shot. A $50 bet on at odds of $500.00 is far riskier than the same stake at odds of $1.32.

Dealing with these problems

To get around these two disadvantages, most punters adopting a level-staking strategy modify things somewhat.

First, the “level-stake” becomes a “level-percentage” of their betting bank. So, for example, rather than committing to placing $50 on every bet, they commit to placing 2% of their betting bank. With a starting bank of $2,500, they start of at $50, but this stake increases with winnings and decreases with losses.

Second, they may use a grading system of stakes depending on the odds they are betting at. For example, an odds-on shot gets 4% of the bank staked, whereas anything bigger than $20.00 only gets 1%, with 2% and 3% used at odds in between.

Your choice as to which staking system you adopt will depend on several factors. If you have a selection approach where you can identify probability with a degree of confidence, and you are psychologically detached from the results of your wagers, then some form of the Kelly Criterion is most likely to maximise your profits in the long-run.

However, if your appreciation of value is more instinctive and approximate, and you struggle to deal emotionally with losing runs, then some form of percentage-level staking is likely a better fit for you.

Whichever method you choose, though, what’s clear is that you should have a betting bank that is kept separate from your day-to-day finances, and that you still need to be able to identify the right bets to make. No amount of mathematical-staking-wizardry will ever change that.