In this article, our senior contributor E.J. Minnis continues his series of "questions and answers" in which he replies to queries sent in by PPM readers. We intend making this series a regular feature in PPM. Please send any questions to: The Editor, PPM, PO Box 551, Dee Why, NSW 2099.

I have decided to take a serious look at staking plans and I am currently reading as much as I can find on the various staking methods available.

I've read a lot about loss-chasing recoupment plans, which mean escalating bet sizes, and this doesn't appear to be a sound approach to me. What are your thoughts?

Someone mentioned to me a staking plan called the Kelly Criterion, which has been around for some time apparently, but it's a method I'd never even heard of until recently - perhaps you could give your opinion on this plan, as well.

One method that really interests me is the flat stake of bank method (where your bet size goes up as the bank rises, with bets going down as the bank does likewise) and wonder if you consider it an ideal method of staking? asks Martin Stevenson via email.

You've raised a few issues, which I will deal with individually before giving you my recommendation.

If your selections lose at level stakes, although not recommended unless the level-stakes loss is minimal, then using a staking plan is the only way of attempting to turn that loss into a profit.

However, I should warn you that any set of selections which loses more than about 5 per cent on turnover at level stakes should be discarded, as the risk of betting on such selections is far too great.

One of the most efficient and high-risk betting approaches of turning losses back into profits are what are generally called "target recoupment staking plans".

"Target" inasmuch as the punter normally sets a predetermined profit or return target that they want to achieve. "Recoupment" inasmuch as the punter increases stakes in accordance with the losses incurred to enable both the losses to be recouped, while at the same time the "target" being achieved.

Such staking plans are not recommended for those who are weak of heart, cannot handle progressive staking (which can get scary at times), or do not have the betting bank to support the staking plan being used. (More on this type of betting in my Punt Master column in the February issue.)

The Kelly Criterion is claimed by many to be the only mathematical true staking plan in existence. Many would also dispute this claim.

In fact, the Kelly Criterion is a money management and clutching system that is designed to maximise the growth of your betting bank over the long term. It gives punters a method of calculating the optimal amount to bet on a horse and the best way to take advantage of overlays.

John Kelly developed it while working at the Bell Telephone Company laboratories in 1956, and it has been mentioned in numerous handicapping and money management books over a period of years since its initial publication.

The strength of this method of betting is that it allows the punter to bet at an optimum level that, in theory, maximises profits.

However, its strength is also its weakness: to bet at an optimum level, the punter not only requires a complete understanding of value and how to correctly price the runners in a race, but also has to wager bets that may be well beyond his personal comfort zone.

There are variations that basically tone down its aggressive approach and one of these variations would be my recommendation to anyone who wishes to adopt the Kelly Criterion as a serious betting method.

The third point you raised was in regard to percentage of bank staking, which is also closely related to the Kelly Criterion.

With percentage of bank staking, it depends on what you are attempting to achieve.

If you have a marginal set of selections that more or less break even, then betting percentage of bank is certainly not the way to go and you will lose more quickly, as you will tend to have your bigger bets on the losers and the smaller bets on the winners.

Also, quite often with this type of staking, you would have to rely on the winners coming in batches, so as to maximise your profits.

If, on the other hand, your selections show a reasonable profit at level stakes, then a percentage of bank method will provide added value and enhance profits.

I would recommend an approach similar to the one that follows:

Betting Bank: $1000

Base Bet: $10 (1% of bank)

Bank Milestone: $1500 (in other words, when you have made $500 profit)

Declare a dividend of 50% of the profit, which equals $250

Reinvest the other $250, which gives you a new betting bank of $1250

New Base Bet: $12.50 (1% of bank)

Continue in this manner of declaring a dividend of $250 and reinvesting the other $250 each time a $500 profit milestone is reached, until such time as your own personal comfort zone has been reached.

Once a new base bet level is reached, do not reduce the bet size until the bank falls below the previous milestone. Such an approach is as good as it gets for those who do not like loss-chasing staking methods.

Terry Schwartz from Melbourne asks a question about prizemoney: My question relates to the average prizemoney figure we often see in many formguides. Is the figure shown an accurate indication of a horse's class compared with that of its rivals? Should we not be concentrating on a figure that indicates prizemoney won, first place only, and then divide that figure by the number of wins attained? Would this be a more accurate indication of the class?

Average prizemoney is only one indicator of class; there are others. Horses at the top or near top of the prizemoney index do win a lot of races, with approximately 65 per cent of winners being in the top four ranked horses by APT (Average Prizemoney Index), while approximately 80 per cent of winners are to be found in the top six ranked horses by API, So don't too easily dismiss this very valuable form indicator.

The average prizemoney index is not a perfect class measurement tool. There can be distortions, generally brought about by horses that earlier in their careers won one or two races carrying high prizemoney, which of course carries on throughout their careers, even though their early potential has not fully blossomed. A better way of using prizemoney as an indicator of class would be to consider what the horse has achieved in either the last year or two years of its career.

The use of a good formguide or database will clearly indicate what a horse has achieved in its last ten or so races, which can be an important indicator of the horse's current ability. So, prizemoney is a good, reliable indicator of class; I use it as part of my analysis process and would encourage others to do so as well.

However, if you do use prizemoney as a indicator of class or ability, then the average prizemoney rankings are a far more important measure than simply the amount of total prizemoney won.

The main issue with concentrating only on winning prizemoney is that it requires the form student to keep records because form providers do not currently present the information in this way, and they are most unlikely to make any such changes in the foreseeable future.

Unless you have some proof otherwise, there is no evidence to suggest that winning prizemoney alone is a better indicator of class or ability than a horse's overall average prizemoney index.

The reason for this is that the ability to win alone is not an indicator of class; many horses do not win, but finish close up in races in a higher class, then step down a grade or two to a lower class and are dominant performers. Your methodology would not take into account this type of performer.

Finally, many people do not realise that a horse's ability is better gauged by their consistency factor (as in the number of first, second and third places they run), as against their number of wins alone.

Tony Markwell from the NSW south coast asks: Form students who do weight ratings use a "qualifying run” as  their base rating. Quite often the "qualifying run" may not be a horse's most recent race, but from a run three or four starts ago. Surely a horse's most recent race should be used as its base rating and not one that may be quite dated. Isn't there a problem with using outdated runs as a base rating?

Also, if the best horse on ratings is too short in the betting market, do you still back it because, irrespective of its price, it is still the best-rated horse in the race?

As a general rule, recent form is the best form to use. But this is not always the case, such as with first-uppers, or horses racing at a more suitable distance than at its previous start. So it is important that the appropriate race is used for a horse's base rating, even if it is sometimes necessary to use a rating that is somewhat dated.

Obviously, value is one of the major issues that make a race a "bettable" race, but that "value" might not be for the win, but could be for the place or in exotic betting.

Quite often when the price of your top-rated selection is unattainable for win betting, value can be found by coupling it with your second, third and fourth rated selections in exacta, quinella or trifecta combinations.

In regard to backing your top-rated selection irrespective of its price, then the simple answer is no value in racing is of utmost importance, more important in fact than backing a winner.

Let me explain a little more in detail: if you repeatedly back horses whose "true" chances are 2/1, but for whom you are getting only 6/4, you will no doubt back plenty of winners, but with a high possibility of eventually going broke.

That is not to say you shouldn't back horses for which odds of 6/4 can be obtained; on the contrary, if the true "value" of such horses is even-money, then obtaining odds of 6/4 will give you a long-term profit.

It is all about value and being able to measure what that "true" value should be. A top-rated selection should never be bet against if it is a genuine favourite. The question with your top-rated selection is to back them only when it is at a value price.

Conversely, a top-rated selection can be bet against if it is not so much a genuine favourite, but simply the most likely winner amongst a number of other possibilities. It may have the best chance, but not such an
undeniable one as to rule out the likelihood of one of your other rated chances winning.

Again, in the latter instance, value should determine which horse or horses should be or not be bet, remembering that value can be obtained in both instances through other types of betting than a simple win-only bet, i.e.
place, exotic or Dutch book betting.

By E.J. Minnis

PRACTICAL PUNTING - JANUARY 2003