By James Willoughby
It is my intention to dedicate this column to the memory of Peter Willett who died this week aged 96. I will never forget being invited to tea at his house in Rotherwick, Hampshire, following a spot-on critique he penned of one of my columns on handicapping in the Racing Post in 2007.
Peter will always be a hero in this household and his books adorn our shelves, most notably his masterpiece, The Classic Racehorse. In a 1984 review of this astounding work, which would be a foundation of any academic study of racing along with Breeding The Racehorse by Tesio and a few others, Jake Spidle of the University of New Mexico wrote the following:
“Peter Willett's book…belongs on the shelves of serious students of modern sport and those of good undergraduate libraries. It is without peer in its field, both in its conceptual approach and in the authority and insight of its author. It is simply amazing how much Willett knows about the history of the breed.”
Peter was an architect of the European Pattern. To say he was interested in the structure of racing and handicapping is a bit like saying Newton was fascinated by forces or that Rutherford was obsessed with atoms. When I visited him, he urged me to have confidence that a lifetime studying the weight-for-age scale or sectional times, or any of other of the academic esoterica related to the sport, would be its own reward. Lulled into a false sense of security, I asked him how he could be so sure. “I am not sure,” he said, “but it isn't given to many to even have the curiosity.”
I can relate this quote with confidence because I wrote it in pencil on the inside-back cover of his book, The Story of Tattersalls. And it is with his words in mind and after his great example that I get to the originally intended start of this article, before hearing the sad news.
Peter cared deeply about the structure of Thoroughbred competition, and so do I. That's why, if I ever had the chance, I would change the handicapping scale on the Flat, because it is patently unfair.
Consider the data in Table 1 concerning Flat handicaps in Britain since 2006. To channel George Orwell, all horses are created equal in handicaps–or so the story goes–but some are more equal than others. See the linear trend in Strike-Rate, Impact Value and Distance To Winner according to a horse's position in the weights (its Official Rating Rank)? Higher Strike-Rates and Impact Values are associated with top weights, while there is an almost perfect correlation between the distance a horse is beaten and its rank in the weights.
Click here for Table 1.
The system is inherently unfair because the exchange-rate between pounds and lengths is wrong. Higher-weighted horses should be conceding more weight to their inferiors, which would equalise the strike-rates and Impact Values in the table. (Impact Value takes into account field-sizes and is scaled where 1.0 is an average Strike-Rate and 2.0, for example, is twice that expected by chance.)
It isn't just the data that points to this but the physics. In Bioenergetics and Racehorse Ratings, engineer Bob Wilkins uses calculus to derive a fundamental scale of weight-for-distance beaten as follows shown in Table 2. This scale is much wider than that employed by handicappers who compile British Horseracing Authority (BHA) ratings, and it prewarns of the effects that using a narrower scale would likely have: higher-weighted horses in handicaps win more often than they should.
Click here for Table 2.
At fault is the scale of 0-140 on which racehorses are rated on the Flat. It is only convention that dictates that jumps horses are rated on a wider scale, which leaves the top National Hunt horses on 175+. Presumably, the root was in the pound-value of handicap weights allotted back in the day: 10st on the Flat is 140lb and 12st 7lb over jumps is 175lb.
Either way, there is no good reason for this evident from statistical analysis. Harmonising the scales would not only remove the egregious effect that is the subject of this article, but also facilitate cross-code comparison.
If racing fans were interested in whether a horse were better over hurdles or on the Flat, the numbers would tell them without the bizarre conversion of adding 40 or 45 to Flat-race ratings to determine whether a horse switching from the Flat to hurdles is well handicapped. It would also resonate with fans to compare and contrast Golden Horn with Sprinter Sacre (BHA 173 over fences), with Golden Horn (BHA 130) more readily than at present.
But, leaving any cascading benefit to one side, the more important reason to expand the Flat scale by 25% is to enable a correction in the exchange-rate between pounds and lengths by the same proportion; in other words, to move the handicapper's calculations more into line with physical reality. This could make a profound difference to the sport, for it would help to tilt handicap wins more across the continuum of power from the powerful stables to those lower down the money tree.
How on earth can anyone justify a system that is predicated on giving all horses an equal chance, when the data proves it does nothing of the sort?
If I could have this discussion with Peter, he might well disagree with me; either way, he might tell me there isn't a chance of the Flat scale expanding to the same as the jumps because of vested interest by powerful connections who have the loudest voices.
But, I have never cared about that last bit. The argument is driven by analysis, by considering two sources of evidence which point the same way. Most of all, however, it is driven by simple “curiosity” in the system. For that and everything else, thank you Peter Willett.
Not a subscriber? Click here to sign up for the daily PDF or alerts.
