Much of what is perceived to be a shift in a stable's fortunes is little more than random variation and what statisticians term "noise"...
Simon Rowlands adopts a novel approach to identifying trainer form and applies it to trainers likely to have runners at the big York meeting...
There is an old saying "there's more than one way to skin a cat".
It is a saying I, as a cat lover and owner, have always found rather distasteful. But what it is meant to describe in metaphorical terms - that different approaches to a problem can each bring a valid solution - is undoubtedly true where horseracing is concerned.
Take so-called "trainer form". There are nearly as many ways of measuring this factor as there are trainers. And, like trainers, some are good and some are useless.
As with other forms of analysis, a crude approach can often give the impression that a trend exists when that impression is just the product of poor measurement and small samples.
Much of what is perceived to be a shift in a stable's fortunes is little more than random variation and what statisticians term "noise".
Even when a stable "is" in-form or out of form, regression to the mean and the corrective effects of the racing calendar (out-of-form horses end up in weaker races, in-form ones end up facing stiffer tasks) and of the betting market can make the "edge" temporary.
That is not to say that trainer form does not exist or may not be useful. Logically, it must be possible for performance to be affected by things like good or bad training practice, by illness, or by some other extraneous element. The difficulty is in isolating it and using it profitably.
Perhaps the biggest problem is the obsession, in some quarters, with measuring trainer form through winners and winners alone. There is only one winner of a race, but there are many shades of performance within that race, and all are pertinent to how a horse, a trainer or a jockey has fared.
In every race, each horse has a price and an expectation of performance, as well as an actual performance that will usually differ from that expectation. Any measure of trainer form - or anything else within racing - which ignores this is probably cutting corners.
I tried to bear all the above in mind when developing a different approach to tackling trainer form this week.
Each horse's odds in a race imply a performance from it in that race. One way of measuring that performance is by percentage of rivals beaten. For instance, a horse with odds of 1.01 has an expectation of beating very nearly 100% of its rivals. At the other end of the spectrum, a horse with odds of 1000 can be expected to beat few of its rivals (though the figure is as much as 16% in reality).
The difference between market-based expectation and racecourse reality reflects how well or otherwise a trainer's charges are performing compared to public perception (as reflected in those horses' odds).
There are a few tricky elements to the calculations, which I can attempt to explain if questions are posted below, but that is the nub of this approach.
The results can be expressed in a number of ways, and I have favoured the factor by which a trainer's horses have beaten other horses compared to market expectation. A figure higher than 1 is good; a figure lower than 1 is not.
With the four-day Yorkshire Ebor Festival almost upon us, I have identified some positives and negatives among trainers who may have runners there.
POSITIVES: George Margarson's horses have beaten 1.30 times as many rivals as could be expected from their Betfair SPs in the last month (28 runners); Tony Coyle is 1.24 (25 runners); Marcus Tregoning is 1.21 (22 runners); Keith Dalgleish is 1.20 (43 runners); and Roger Varian is 1.19 (69 runners).
In addition, while not scoring quite so highly, the bigger battalions of Sir Michael Stoute (1.14, 64 runners) and Richard Fahey (1.09, 204 runners) have overperformed and may continue to do so.
NEGATIVES: William Muir is 0.81 (46 runners); Amanda Perrett is 0.84 (48 runners); Jim Goldie is 0.85 (73 runners); and Michael Bell is 0.87 (66 runners).
William Jarvis has had only 19 runners in the period under review - a rather small sample - but his figure of 0.51 (his horses have beaten roughly half the number of rivals predicted by the market) was by some way the worst of all trainers surveyed.
Most major stables were neutral, the market, unsurprisingly, neither over nor underestimated the horses' abilities overall.
The market will, in due course, adjust to the others as well. But a lot of the information which drives the above figures is "hidden" from those fixated with winners, or even with winners and placed horses. So there may be a lag which punters can turn to their advantage in the shorter-term.
If nothing else, it is a different way of skinning this particular cat, and a different approach sometimes pays dividends.
