A problem arises when we find we have too many bets for the day. The first thing we do is to go back over our selections to see which ones we should drop. We also suspect that whichever ones we leave out are going to romp in at ridiculous odds for no other reason than we decided to leave them out of our betting activity for the day.

Horses can be graded in terms of their chalices of winning a race and the method used to grade your selections call be as simple or as complicated as you like. The plan of attack here is to acknowledge that on some Saturdays there will be more selections than you really want. Also, on other days there will possibly be too few, although I do not see this as a problem at all.

Let's face it. If you only have two bets for the day and one of them wills, you will be happy enough. Alternatively, if both lose, you have not lost very much so you live to fight another day, okay.

Should the situation arise whereby you have something like 13 selections for a day's betting, then you have a problem. The break-even odds will probably not be attainable and, whilst you might pick three winners, the chances are that you might still come out losing on the day.

To get 20 per cent on turnover, you need the win dividends to be about $15.60, which might be a bit of a problem. I would prefer three bets, with one winner at, say, $4, which is far easier to achieve.

If we are all agreed, which never happens by the way, that dumping some of these selections with a view to reducing the number of bets for the day is desirable, or a better plan is required to allow for our superstitious natures, then here is what I would suggest you consider.

First the simple approach. Look at your selection's current win ratios and break them up in descending order. This now allows you to divide your selections into three grades or types. You could also use the pre-post or actual betting market to do the same thing.

The next step is to acknowledge that the best three or four deserve the most money bet on them, the middle section should have less and the third group are really i token or dabble bet, if you are being honest with yourself. This last set is going to let you down in the long term for sure.

Now the bet ratio is three units on the best, two units on the middle section arid one unit on the last set.

Got the idea.

So instead of having three units on all 13 selections, which totals 39 units, we end up with something like this.

The first group of, say, four horses gets three units bet on them, the second group, say another four horses. gets two units bet on them, and those in the last group get one unit each. The total outlay here is 25 units. I have saved 14 units outlay if there was a complete wipeout for the day. Picking two winners from the first group and one from the second group will give me a much better chance of making a profit for the day.

What I am talking about here primarily is preserving the betting bank and still allowing you to bet on all your final selections. Any system that consistently puts you in this sort of position of having too many final selections has not been refined enough, in my opinion.

You can't really expect to pick winners each and every week that are going to pay you win dividends in excess of $13, can you? It can and will happen on occasions but over a year you are bound to be losing substantially.

Another way you cart end up with the same problem is if you are following more than one system. Each system call throw up its three or four selections but if you were following three systems you could end up with the same problem.

Firstly, if more than one system picks the same horse, do not consider this a double bet, the horse does not know that it has become a system selection (twice) so its chances of winning are the same.

That helps a little but then you have to go back and start grading all the selections to arrive at the better-class selections. In this way you can bet more intelligently and know what you are doing before the event when your head is a lot clearer.

If you want to use a more sophisticated method of grading the selections, then race ratings would have to be the way to go. These are available from many sources now and will definitely tell you who the better-class horses are.

Using the same principles as above, you can then bet accordingly. Some people like to use average prizemoney as a measure of class, which is also okay. Whichever method you prefer. I would advise you to be consistent arid stick with the same one.

• The Keith Roth File appears in PPM each issue. Keith covers many areas of betting, race selection and staking.

By Keith Roth