Is there any value in progression-style staking plans? Can you really beat level stakes betting? In this article, new contributor Arthur York poses what he considers are valid questions about progressive staking.

Let's begin by saying there is no perfect method of staking.

There can be no perfect staking system. I have always believed that the method hasn't yet been formulated which will beat the simple approach of betting level stakes.

I maintain that most staking plans are based upon a simple failure to get to grips with the reality of 'sequences'. Too many people believe in that old fallacy of the so-called 'law of averages'.

Only the other day I was scanning through an overseas publication and noticed an erudite article on this very subject, the writer pointing out that if a jockey has ridden 50 losers in a row he is not one whit more likely to ride a winner next time he rides. If the favourite has failed in the first seven races of the day, that does not mean the last favourite is any more or less likely to win.

The law of averages, this writer said, made people believe that something must happen because of what has happened before. But there is no such thing as the law of averages. Says the overseas writer: "If you toss a coin 100 times it's an even chance that you will get less than 57.75% of either heads or tails; if you toss it 1,000 times it's an even chance that you will get less than 52.133% heads or tails; and if you toss it a million times the odds are that you will not get more than 50.0674% heads or tails."

The more trials you make of a chance event the nearer the result approaches the true theoretical probability. But now the staking systems as they apply to horseracing, or dogs and harness racing. Firstly, we can say that any system that asks you to wait until a certain sequence of results has happened, and then step in and start betting, can be chucked in the rubbish bin immediately.

Most progressions are based on increasing your stakes after a loser. The basic plan is more or less the following: 1 unit for the first bet, increase 1 unit after each loser, after a winner revert to 1 unit and repeat.

My point is that all systems calling for increasing stakes after losers depend for their success upon the sequence in which the winners and losers fall. If the winners are well spaced out, with a lot of losers preceding each winner, they are successful, but if the winners occur consecutively, or bunched together in what I like to call 'cluster bombs', then they are unsuccessful.

Let's look at a series of 20 bets, made up of 15 losers and 5 winners priced at, say, evens, 2/1, 3/1, 4/1 and 5/1. If you bet this series of bets at level stakes the punter would be left all square no matter which way the winning or losing sequences happened.

But using the staking idea I mentioned earlier (increasing 1 unit after a loser and reverting to the minimum 1 unit after a winner) the profit or loss depends entirely on the way you arrange the winners and losers.

Let's look at sequence A: LLLL3 / 1, LLL5 / 1, LL2 / 1, LLL4 / 1, LLL evens.

The result of this sequence is splendid: A profit of 30 units. But let's look at the same horses when the winners are bunched in sequence B: LLLLLLLL evens, 5 /1, 4 / 1, L 3 / 1, 2/1, LLLLLL.

Ah, what have we here? A loss of 34 units!

Now let's turn to decreasing your stakes after a loser. The basic approach with this is you bet in 1 unit stakes until you strike a winner, then jump the stake to 4 units (for example); decrease 1 unit after each loser until you reach the minimum 1 unit stake, with which you continue until another winner jumps it to 4 units again; always bet in 4 units after a winner.

The reasoning behind this approach is that when you strike 'form' you will frequently be using 4, 3 or 2 unit bets but when out of touch you'll only be betting 1 unit. Now, the plan can be a winner when the winners are together in 'cluster bombs' but it will fail if the winners are spaced well apart.

Let's look at sequence A for this idea: LLLL 3/1, LLL 5/1, LL 2/1, LLL 4/1, LLL evens.

This ends up with a loss of 21 units, but now let's look at sequence B: LLLLLLLL evens, 5/1,4/1, L 3/1, 2/1, LLLLLL.

Here we arrive at a profit of 30 units. What we can deduce from all this is that the two staking plans contradict each other. They cannot both be right, for if the sequence of results over a long period suits the first plan it won't suit the other one, and vice-versa.

In real life (!), everyone gets sequences of results which are, in the long term, of both kinds and, as a result, both plans are neither better nor worse than ordinary level stakes betting.

There are lots of variants upon them but in the end they all come to the same thing. Remember, too, that the test of any staking system is to find the true result: That is, divide the final profit, or loss, by the average stake used.

Bearing in mind, then, that I advocate the easy method of putting the same amount of money on each selection, let's look at the potential for losing runs. My figures show that if you were lucky enough to have a selection
method that produced 50 per cent winners (tell me about it, PLEASE!!) you could expect a run of 10 straight losers after an average 2048 bets, and a run of 15 straight losers after an average of 65,536 bets.

But if your plan only gets 20 per cent winners, then a run of 10 straight losers can be expected after an average 47 bets, and a run of 15 losers after an average 142 bets. If your plan gets only 16.6 per cent winners, then the 10 straight losers could be expected after an average 37 bets.

So how large does your betting bank have to be? For a bank to be considered satisfactory it must be large enough to withstand the worst losing sequence that can arrive. This depends not only on the percentage of winners but on their average price.

A plan which gets 50 per cent winners requires a smaller bank than one which hits, say, 25 per cent winners. What really matters is the actual 'rate of profit' achieved, not the probability of a given number of straight losers.

Those punters astute enough to achieve a high profit rate with 25 per cent winners could operate on a 10 per cent of bank approach. But I would go further and recommend a 20 per cent approach - provided, of course, that your one winner in four is going to return a high rate of profit.


  • For any punter to win at racing he must discipline himself to have a bet only when he considers the odds to be attractive - that is, in his favour.
  • Your general level of betting has to be governed by your financial resources.
  • Remember that all staking plans rely for their success on winners and losers occurring in the kind of sequence which suits their rules.
  • Level stakes betting allows you to remain calm in any situation. You do not have to cope with the mental strain of increasing bets.
  • Check out how your bets have fallen in sequences - if you are planning a staking progression then rearrange your winners and losers and cheek out what happens!
  • If you hate level stakes with a vengeance, and want more thrills, then bet according to your selection's true chance. If you feel a horse is a 6/4 chance, then back it as such and ensure you get an overlay.
  • Bet when the odds are in your favour. Otherwise stay out.
  • Get yourself a decent bank, assess what rate of profit you can achieve, and bet accordingly.
  • Decide if you're a cautious bettor or a risk-taker. If the latter, then you will always bet in higher amounts than the cautious punter.

By Arthur York