Conventional race analysis – a four step approach
In the last article, I referred to conventional form-based race analysis and subsequently received a few emails requesting more detail about this method. Essentially, there is no single form-based method, in fact there are probably as many approaches to form-based race analysis as there are racing systems, and each bettor would have his/her preferred technique or range of techniques. However, in the following sections I have presented one interpretation of this approach to winner-finding which I apply to flat handicaps.
The first step when applying a conventional race analysis method is very straightforward: select an appropriate race or group of races to examine. For my approach to flat handicaps this means focusing on just those races with ten or fewer runners. Unless there was a significant draw bias I would not even consider a large field handicap for betting or tipping purposes.
Given the amount of racing currently staged each day there should be plenty of suitable races in which to bet without having to attempt to find a value bet in races with far too many runners to make them viable.
Step two is to check the ratings. I use speed ratings and would normally focus of the top-rated runner. However, speed ratings are notoriously volatile, so it’s best to find a runner with a good ratings profile as opposed to a single good figure. Exceptional performances can be the result of a variety of reasons, including weather patterns and field sizes.
Consequently, a single good performance, which puts the runner at the head of the ratings for a race, is not necessarily reliable. Far better is to find a top-rated runner which has one or more other ratings close to his/her best figure providing a degree of support for the high rating.
As an example, I would consider a horse top-rated with a figure of 72 unreliable if the ratings profile was as follows: 56 50 72 45 58. Alternatively, a profile of 56 55 65 72 70 would be more acceptable since the 70 and 65 support the top figure of 72. The order of the ratings is also important.
If the top-rated figure has not been achieved for some time, say three or more runs, then I would mark the ratings profile as poor; ideally the highest figure would have been recorded on the runner’s last, or penultimate, start. The ideal ratings profile is one where the ratings steadily increase showing a clearly improving trend.
If the ratings profile for the top-rated runner looked dubious, I would then look to the next highest rated horse, or maybe even the third highest rated runner if it had a good profile. If none of the top three rated runners possess a good ratings profile, then I would move onto the next race.
Once a horse with a good ratings profile has been identified the next step is to determine whether there are any obvious reasons it will not win. Essentially, I check for six factors when validating a selection. First, is course absence. If the runner has been off the track for over 80 days, and he/she has either performed poorly after a break in the past, or the trainer has a poor record with horses returning from a break compared to his/her other runners, then I would pass over the horse.
The second factor is going suitability. A horse would be removed if he/she had only ever run poorly on similar going and there was clear evidence that the ground was the cause of the below par effort(s). The same would apply to the race distance.
The fourth factor is recent trainer form. It is reasonable to expect a trainer to have an 8-10% win rate in flat races given the average field size, so if his/her recent win rate was below 6% I would rule the horse out.
The draw is another important attribute to consider on some tracks. However, in recent years the number of biased tracks appears to have been reduced so there are only a few key courses at which I would apply this elimination rule, Chester, for example would be a classic case.
The sixth factor is the jockey. The rule I use is very simplistic: I would eliminate the horse if the jockey claimed 7lbs or more in weight. Many analysts would find this rule too basic and I am sure more sophisticated rules would perform better, but this would entail a great deal of additional work analysing performances of jockeys and I am not sure the benefit would outweigh the effort. If the selection failed one or more of these six checks then I would repeat the ratings analysis in step two in order to find an alternative runner.
The fourth and final step is to find a reason, other than the ratings, to opt for the horse. The results of handicap races are difficult to forecast because the Official Handicappers are good judges of races and set the handicap marks appropriately. Basing a selection solely on the ratings assumes that the set of figures used is better than those determined by the Official Handicapper which, in the long run, is probably not true.
Obviously in cases where the runner has an improving ratings profile it might be true that these are under-assessed by the BHA, but when risking your own money, or worse still advising others to risk money on your prediction, it is better to have other evidence to suggest the horse will run better in the race than he/she has run previously.
A return to optimal race conditions in terms of going and distance, as well as course, is a positive, but only if the horse has been assessed on performances on recent sub-optimal conditions. In this case a return to better conditions should bring about improvement compared to recent runs.
Excellent recent trainer form is another factor which can make the difference. A horse from a stable which has a recent win rate of 20% or more is likely to find some improvement which, in addition to a good ratings profile, makes him/her a viable selection.
Other positives are a quick return to the track, horses running in all age handicaps after a break of no more than seven days win at a rate of 14% compared to the 10% average.
Finally, and most important to me, is whether the horse is improving. The ratings can show improvement, but another factor is the number of times the horse has raced. Fewer races means more chance for improvement. A change of race distance (see previous article) or trainer can also induce improvement and in many cases can be viewed as a positive.
If the horse passes these four steps then it can be considered a viable selection, though it must be noted that this method does not provide a mechanistic way of determining a fair price for the horse. However, if you analyse many races this way it is possible to gain a feel for a fair price. Whist this sounds dubious in the extreme, it is surprising how accurate you can become over time by repeatedly applying the same approach.
Fortunately, if it is not possible to gauge a fair price from the analysis all is not lost. Successful systems followers rarely incorporate fair price in their methods: the system employed determines a selection which possesses a set of attributes that in the past have isolated a set of runners that have been profitable to follow, and therefore is a value selection regardless of the price on the day.
Essentially, these four steps form their own system and can be applied as such, so the price available then becomes redundant, as it does with normal systems use, providing that this approach is found to be profitable in the long run. So it is critical to record the results of previous analyses to see whether the method works for you.
I can only provide data for the last 400 races I analysed for ATR using this method without applying any price filter, and these races returned a win rate of 37% and a level stakes profit at Bookmakers’ starting price of 5p/£ which equated to 15p/£ at off time exchange prices. This suggests there is potential in this approach and, hopefully, if you decide to base an analysis method on these four steps your profit will be even higher.