James Willoughby Op/Ed: A Numbers Game

Racing. It’s a numbers game. Take the impending retirement of one of British and Irish jumps racing’s biggest heroes, A P McCoy: “More than 4,300 wins,” “About to claim his 20th straight championship,” “Winner of 289 races in the 2000/2001 season,” the stories blared. 

These are the numerical dimensions of McCoy’s greatness, the accumulation of 20 years of drive, determination and deserved opportunity. But they are only so-called counting statistics, which rely on an understanding of the underlying context, the difficulty inherent in their achievement. In 1989, the U.S. jockey Kent Desormeaux rode 598 winners, while his contemporary Russell Baze has 10,533 career wins. Heck, there are nine million bicycles in Beijing–and that’s a fact. 

Statistics without context–or without, more properly, population parameters–are like trifles without sponge: a mess of ingredients with nothing on which to reside; just baseless, abstract mush. You’d be far better off consuming something else.

Of course, the melange of McCoy numbers does no harm. It’s not like governments abusing economic data or scientists funded by drug companies massaging medical research. And, in any case, there is no ambiguity in the conclusion: whether you appreciate McCoy by the empirical or the aesthetic, he’s undoubtedly one of the toughest, most successful sportsmen of all time. 

The problem is the conflation of the concept of greatness with accumulation. In many cases, as with McCoy’s, racking up huge win totals can be taken as substantive evidence of the more important notion of efficiency, but in many other cases opportunity must first be discounted from achievement to understand who really is the best at what they do. 
In economics, it is a central tenet that efficiency is best considered at the margin. In other words, what is the effect of a particular person or strategy, holding all other things the same. 

To evaluate jockeys properly, we should not just count up their wins but ask: how many winners might an average or so-called replacement-level jockey (one typical of those readily available for employment) have managed in their place? 

It was exactly the ability to ask–and answer–this question that enabled the sport of baseball to make such dramatic strides in self-analysis. Brought to wider attention by the book and film Moneyball–perhaps a kind of Fifty Shades of Grey (Matter) for those gagging for more intellectual stimulation–the analytics movement empowered those teams without the greatest financial muscle (ie. the Oakland Athletics) to run up bigger regular-season win totals than those with more resources. Is a similar enlightenment possible in racing? 

The answer is a definite ‘yes’. It’s already happening. There are smart people all over the sport who know how to exploit the market inefficiencies caused by the many traditional beliefs that hold in the most traditional of sports. First there were bettors, newly armed with databases or prepared to hire coders to find systematic edges in betting markets, notably Zeljko Ranogajec in Australia. More subtly, stud farms and racing operations all round the world are waking up to the advantages of analytics, which the rest of the commercial world previously assumed. 

Not every brilliant jumps jockey has the opportunity to ride 200 winners in a season, and not everyone who rides 200 winners in a season is a brilliant jumps jockey. If it’s just for the sake of a water cooler or bar room argument, it doesn’t really matter who we credit subjectively with various qualities. But it absolutely does matter if we are going to rely on their services to gain an advantage over the competition, especially if we have to pay more for past glories that cannot be sustained. 

There are few areas in which a Moneyball revolution in racing could provoke more changes than the sales ring. At a basic level, the data generated by racehorses is misappropriated wildly by the market with the excuse that bloodstock buyers are more motivated by chasing a dream than by efficiency. 

Let’s take stallion statistics, for example. There are some impressive, well thought out and well executed methods for expressing the results of the offspring of sires (and, by extension, their potency as progenitors of Thoroughbred talent) but in Europe it is rare to hear the quality of a stallion’s book of mares discounted from his results. (Note that this is not so much the case in the U.S., thanks to intelligent metrics such as CI and CPI. Still, we are a long way here from stallion statistics being presented in the same rigorous, objective manner). 

It is highly likely there is at least an equally strong association between the quality of a mare and her offspring as that between sire and foal, so, back to the economics 101 classroom, if we manipulate the book of mares for all stallions to median quality, we might find the established order of stallions is very different. You might still believe that the cream would rise to the top but, from the standpoint of making trifles, it could now be the sherry or the custard. 

To many in the sport, mathematical analysis of the port at this level arouses as much suspicion as it once did in the front offices of baseball teams. And, yes, it is not facile to reach objective truth about the sport through numbers, but neither was it easy getting to the Moon–and we managed that in 1969. 

The difference between rigorous analysis in horse racing and space exploration, however, is motivation. It strikes me through bitter experience that upsetting the status quo publicly in racing benefits few, while sticking the Stars and Stripes in the nearest rock to Earth was bound to be more warmly received. 

Behind the scenes, however, is a different matter. Racing operations all over the world are increasingly recruiting and training young people with decorated academic backgrounds in business, finance and economics to evaluate the data produced by the sport. Many small steps could become giant leaps in the next few years. 

To misquote John F Kennedy (the U.S. president, not my ante-post Classic wager trained by Aidan O’Brien): “No racing or breeding operation which expects to be the leader of others can expect to stay behind in this race for data. We choose analytics not because they are easy, but because they are hard.”