'Why do horses win? Why do horses lose? Which horses are more likely to win? What factors can we use to pinpoint these winners before the race? Yes, these are just a few of the host of riddles which confront punters every racing day.

Between 1984 and 1988, I conducted a survey of how favourites fared in races. My study revealed a win strike of 33.7 per cent from 2055 races-which, put another way, means that 66.3 per cent of favourites lost! But the survey threw out some very interesting points in regard to the performances of favourites in different races.

For instance, it showed that a favourite wins at least one race virtually every day and, on average, will win two or three races per meeting-but favourites virtually never win more than four races on a card.

Maybe, then, you can properly assume that once four favourites have won then that's it for the day-you can discard them for the rest of the programme (assuming there are other races left to be run). Most times you will be absolutely correct in adopting this philosophy. Another angle we can assume-from my survey, anyway-is that if no favourite has won in the first four races then we can expect two, possibly three, favourites to score at the remainder of the meeting.

Now nobody claims that the law of averages works infallibly, but the evidence of the figures is overwhelming. But week after week the figures of two, three and four favourites were recorded. If, then, five or six races pass without a winning favourite we are looking good to get a winner from the last two or three races. Why? Only nine times in five years (3.4 per cent) did meetings finish without a favourite winning.

Sometimes, of course, we'll get the 'wacky winners', those horses that defy all the percentages. This is only to be expected. If racing did not have its surprises we would have no racing.

Let's now look at some other interesting figures relating to where winners come from and how we can spot them. Some years ago in P.P.M., Statsman produced his own set of survey figures relating to many different aspects of winner-finding. His statistics were compiled over an eight-year period, and comprised barrier positions, tipster poll positions, last-start placings, days since last start, winning TAB numbers and position in the morning newspaper betting markets.

Statsman has told me he has continued to keep the statistics up to date and reports there has been very little changeso the 1986 statistics remain totally valid. Any punter seeking to contract the percentages stacked against him would do well to closely study Statsman's findings, because by doing so the punter can often determine which horse, out of several possibilites, is the most likely to win, based on what has happened in the past.

Barrier positions (after scratchings) are important because the inside five barriers provide just over 42 per cent of all winners, with 1 and 2 being best with 10 per cent and 9.3 per cent respectively. Horses at the top of newspaper tipping polls (like those in the Sydney Morning Herald, Melbourne Sun, etc.) win almost 27 per cent of the time, with the second choice winning 17.2 per cent and the third choice getting up 12.5 per cent of the time.

So, the first three horses in the polls will win 56.5 per cent of races. Not a bad statistic to consider when you are examining prospects for a race. Horses at the top of the polls, by the way, run a placing (1 /2/3) 59 per cent of the time.

As far as TAB numbers are concerned, No. 1 wins 19 per cent of races, while No. 2 wins another 12.6 per cent and No. 3 another 13 per cent, giving those three numbers a grand 44.6 per cent of all races-yet another very interesting statistic, which an astute punter can use to his advantage in the long term.

I know many punters take a great deal of notice of the morning newspaper betting markets. Statsman's figures show that you can expect the pre-post favourite to win 31 per cent of races, and the second favourite 20.5 per cent, giving those two more than 51 per cent of races.

Now to a more recent study, this time conducted by a part-time researcher for P.P.M., and embracing some 800 races. What our researcher did was to look at 800 second and third placegetters (last start) who went on to win (favourites were eliminated, as were some horses who were well-backed because of top-line jockey or trainer influence).

The range of odds covered from 4/1 upwards, and the results showed that of the 800 winners, 270 were at 4/1 or 9/2, 220 at 5/1 and 6/1, 200 between 7/1 and 8/1 and the remainder at 9/1 or longer.

Significantly, a high proportion of the winners had finished UNPLACED at their second-last starts (prior to finishing 2nd or 3rd at their last start). Horses beaten narrowly (up to 2.75 lengths) won many more races than those horses which had been beaten three to six lengths.

Surprisingly, a horse's career consistency had no marked effect on the number of winners. Horses with low win percentages managed to win just as many races as those with a strike rate of 20 per cent and more. A majority of the winners had lasted raced within the three weeks preceding (but it must be pointed put that the researcher discarded any horses which had not raced for 30 days or more).

The sum total of our researcher's probe was that horses at 8/1, 10/1 and 12/1 have just as much handicapping 'thump' as those horses at much lower prices (4/1, 5/1). So these, then, are the horses being ignored by most punters.

What it means is that you could be on to a real winner-finder by looking for horses at attractive odds (8/1 upwards) which finished 2nd or 3rd last start, especially if they had their last starts within the previous 15 days, and if they were unplaced at the second-last start, and if they finished within 2.75 lengths of the winner at their last start.

Tie this research in with Statsman's statistics and you could then come up with a string of positive factors which could lead you to winner after winner.

Perhaps one of the most important statistics you should consider is that relating to the number of days since a horse last started. Statsman's work shows that more than 85 per cent of all winners had their last start within the previous 21 days. Even more significantly, horses which haven't raced for 29 days or more win only a shade over 9 per cent of races.

The warning sign, then, is to look very carefully - and suspiciously-at any horse which hasn't raced in the previous three weeks. The percentages are stacked against these runners.

A further aspect of all this statistical stuff relates to actual prices. Can you back short-priced horses and make a profit? The simple answer is 'No'-unless you can eliminate some of the losers.

In his book Gambling Into The 90s, the Sydney mathematician Hans Eisler tackles this particular aspect of punting in enormous detail. You might say to yourself that you can win money backing 6/4 chances. This gives you an expected success rate of 4 in 10 (40 per cent). But, says Eisler, the actual success rate will vary between 310 and 394 in every 1000 races, leaving you between 1.5 cents and 22.5 cents out of pocket on every dollar invested.

You can, however, expect 832 of these 6/4 chances to finish in the first three placings in every 1000 races. Which is something to think about, especially if you were able to successfully weed out the false 6/4 chances! If you can do that, the prospects of long-term profit rise with every 'falsie' you nail.

Finally, a few other 'wacky' facts that you may be interested in as part and parcel of your punting armory. I dare say you have often been in the situation where you have backed a winner and then a protest is lodged? Of course you have. I know I've been in the predicament many times, sometimes wishing the protest would fail, other times praying it would succeed.

But what are the chances of a winner being disqualified on protest? My figures say it's 771/1! So if your winner is under the gun don't worry too much. The odds about a dead-heat are around 211 /1.

The odds are 27/1 against a TAB Daily Double paying more than $100 for a 50 cent bet. In a test of 6,000 races-carried out in the early '80s-it was found that only 25 were won by horses at 50/1 and longer-less than half of one per cent!

A friend of mine-a "chap i meet regularly at the races-once asked me to explain the ins and cuts of each-way betting to him. He couldn't really grasp the shortening of the odds that occur with an each-way bet.

He didn't realise-and I'm sure many punters don't-that if you back a horse $1 each-way at 4/1 and it wins you actually have taken odds of 5/2. You see, if you had put your $2 on for a win you would have collected $10. The each-way bet returns you only $5 for the $2, giving you a final collect of $7, or 5/2. The 'safety' place component of the bet, while OK if the horse is 2nd or 3rd, drags down the overall return if the horse wins.

Statistics, as you can see, are interesting and there is no doubt they can play a key role in helping a punter to avoid falling into the mistake of bucking the percentages. Set your mind to thinking about the figures I've unveiled here and I'm confident they will enable you to at least polish up your form study, as well as provide you with much 'food for thought' .

By Richard Hartley Jnr