There are thousands of punters around Australia who would like to play their own individual systems - but lack the information required to check back on past results to ascertain if what they have worked out actually has a snowball's chance in hell of surviving.

We know from the amount of letters received at P.P.M., that many readers have any number of bright ideas; what they don't have is the information at their fingertips to do all the necessary checking. With systems, and with any idea in racing pertaining to selection, you have to rely a great deal on the past to accurately predict what might happen in the future.

These days, fortunately, there are advanced computers and their associated software available to those people who have the money, the inclination and then the ideas to take advantage of them. Because of P.P.M's computerised set-up we are able to look at various angles to see if they have worked, and are likely to work in the future.

Our computer expert, Neale Yardley, began work last year on the exacting task of feeding various statistics into the computers to see what they could tell him. He could do this because of his access to the vast databases of Australian Associated Press and George Tafe. Contained in these databases are tens of thousands of race results, complete with everything you need-margins, weights, jockeys, price, weight and class ratings, barrier draws. etc.

Any horse's form can be traced back for years. Neale's first venture was to prepare a special feature for the 1991 P.P.M. Annual. Now he is producing a series of articles for the monthly P.P.M., in which he will examine various aspects of form study, with a view to testing any number of theories regarding key angles.

Neale's research will be of immense benefit to punters who have sought the answers to some of these angles for a long time. I know I found the information absolutely rivetting and indispensable, really, in looking at form.

If you have a system idea you would like to see tested, drop me a line at P.P.M. and, if it's suitable and able to be checked, we'll let Neale have a look at it by putting it through the computer.

The following is Neale's first article in this intriguing P.P.M. series.

Racing statistics are extremely important to punters. One of the best things they can be used for is to help punters turn their punting losses into profits.

Unfortunately detailed racing statistics are not often made available to the average punter. Expensive computer form databases and statistical analysis programs are still only the province of a handful of Australia's top form experts. It is for this very reason that we have decided to provide you with access to such research through this new series.

Many of you will recall how in Breakthrough I presented a range of statistics on the percentage of starters in races that won or were placed at their last start, and (more importantly) the percentage in each group that went on to win their next race.

The main conclusion to be drawn from the statistics I compiled for Breakthrough was that last-start winners and placegetters were much more reliable betting propositions than horses that were unplaced at their last start.

Fortunately I can tell you this without giving away too much about the rules The Optimist derived for Breakthrough. Needless to say Breakthrough does have a rule about considering last-start winners.

It may come as no surprise to you that many successful systems start with a rule requiring you to consider only last-start winners. I used such a rule to develop the system in my "What If?" article in the 1991 Practical Punting Annual. What I am not going to tell you is what the additional rules were that I used to end up with a system producing a 23% strike rate and 27% profit. (It should go without saying that if you haven't ordered your copy of the Annual yet, you should do so now.)

Rather than developing a system in this article like I did in the Annual, I am going to concentrate on looking at statistics. In this and future articles in this series I will be drawing on race results for the racing season August 1, 1989 to July 31, 1990. My computer analysed nearly 50,000 results during this period from just under 5,000 metropolitan and provincial races in Victoria, New South Wales, Queensland and South Australia. (Note that these figures suggest an average of 10 horses in a race-a figure I will be using shortly.)

One of the first things I set my computer to look at was all homes that were first, second or third at their last start. I found that 12-3% of these went on to win their next race.

If we consider the last-start winners rule that we mentioned above, the strike rate improves to 14.1 %.

Before moving on you should note that these strike rates have been calculated for metropolitan tracks only. In practical terms this means for example that we look at all last-start winners that are racing at metropolitan tracks. Note that this does not mean that non-metropolitan results are not considered-they are to the extent that non-metropolitan form leads into metropolitan form.

In case you are wondering whether these figures are good or bad you should remember that on average there are about 10 horses in a race so that all other things being equal a horse should have one chance out of 10 or a 10% chance of winning its next race. This is the benchmark strike rate we need to improve upon and clearly the figures just given are better than this.

Now you should understand why last start placegetters, and in particular last start winners, are much better betting propositions than unplaced horses. In fact if we look at horses that were unplaced at their last start we find that only around 6% of them win their next race.

For those of you who might be interested we will now take a quick look at some last-start winner statistics for Sydney and Melbourne.

In Sydney, 13.7% of last-start winners went on to win their next race.

In Melbourne, 12.3% of last-start winners went on to win their next race.

As you can see the strike rate is marginally better in Sydney. Such a difference however is not necessarily significant and we will have to wait until the end of the current racing season before we can confirm whether a trend exists.

You should note that if strike rates turn out to be consistently better in Sydney than Melbourne it would not automatically mean that Sydney horses were more consistent than those racing in Melbourne. All you would need is for the average number of starters in Sydney races to be one or two less than Melbourne races and you would end up with the same difference in strike rates.

The strike rates mentioned so far were achieved by backing last-start winners/ placegetters at metropolitan tracks irrespective of the day of the week or venue of a horse's last start. In an attempt to confine oneself to better quality horses it is not uncommon to consider Saturday metropolitan racing only.

My computer told me that if I confined the analysis to Saturday metropolitan racing only, the strike rate for last-start winners improves to 15%. Interestingly enough the figures are 15.3% and 12.7% respectively for Sydney and Melbourne (the margin in favour of Sydney as before although slightly greater this time).

We will next consider last-start city winners, namely horses whose last-start win was at a metropolitan track. This is another useful way to confine oneself to the better quality horses.

My computer told me that 17.8% of last-start city winners went on to win their next race (18.1% for Saturday races only). As you can see the strike rate has suddenly increased by nearly 4% just as a result of further confining oneself to last start city winners.

I was going to devote a fair bit of space to days since last-start statistics in this article but decided to commence with the more important last-start winner's statistics.

It turns out that days since last start only become a significant disadvantage for horses that haven't raced for more than a month. I found for example that of horses that had last raced within 28 days, 9% went on to win their next race. For horses that had last raced within 14 days, 10% went on to win their next race.

I suggest you use days since last start in conjunction with other selection methods rather than in isolation. By itself, days since last start produces no better than the 10% average strike rate. In conjunction with another selection rule however, strike rates can be improved.

For example, last-start city winners that last raced within 0-14 days we find th-at such horses go on to win 19% of races nearly one per cent better than the strike rate achieved when considering all last start city winners. This is not surprising as although last-start winners were likely to be at their peak at their last start, they are only likely to retain this condition if they had their last start fairly recently.

Lees summarise our findings.
RANDOMLY PICKED HORSES should win around 10% of races.
LAST-START PLACEGETTERS should win around 12% of races.
LAST-START WINNERS should win around 14% of races.
LAST-START WINNERS AT SATURDAY MEETINGS should win around 15% of races.
LAST-START CITY WINNERS should win around 18% of races.
LAST-START CITY WINNERS THAT LAST RACED WITHIN 0-14 DAYS should win around 19% of races.

In concluding you should note that by applying some fundamental criteria regarding a horse's last start we have almost doubled the average strike rate of 10%.

At the outset I mentioned how racing statistics can be used to help improve ones profit. Readers of my Annual article will know that the way to do this is by developing a system. The more statistics you have, the more system rules you can come up with. Provided your system rules are aimed at increasing the strike rate by weeding out losers and holding on to winners, you will find that the strike rate continues to improve over and above the average 10% benchmark figure.

I indicated in my Annual article that as the strike rate of a system improves so too does the profitability. Mat must be stressed is that if you are after a profitable system, you must first look for a method that produces a high strike rate. If you ignore the strike rate and just look for rules that produce good-priced winners then you are likely to end up with too few selections and a low strike rate. While such systems can be profitable, they can very quickly become non-profitable simply by the absence of one long-priced winner.

In future parts of this series; we will continue to look at statistics that improve upon the average 10% strike rate. Invariably we will be doing this by considering statistics that concentrate on horses that are fit, in form and ready to win.

While we will be concentrating on strike rates in future articles, readers can read a more detailed analysis about the profitability of individual statistics in my monthly newsletter Punter's PC Digest.

Next month we will continue to analyse last-start winners but in addition to looking at their strike rate for winning races we will also look at their strike rates for coming second and third in subsequent races.

Click here to read Part 2.
Click here to read Part 3.
Click here to read Part 4.
Click here to read Part 5.
Click here to read Part 6.

By Neale Yardley