As mentioned in the previous article, speed figures are only one part of the handicapping toolbox, but they are a large part, because they quantify other variables to simply; how fast the race was run.
Like everything based in science and statistics the strength of the figure comes from the amount of data that is available in the sample size. A computer program that can pour through all the past races at Saratoga since 1990, will give you a very solid foundation for analysis, as opposed to looking at the results of just a week or two.
Yet there remains a lot of controversy when it comes to speed figures. Routed in skepticism and misunderstanding, most of this is from the figure makers concession that they must use a certain amount of subjectivity when coming up with a variant for the race. This subjectivity comes into play particularly when tracks conditions have changes within a short amount of time. There is a subtly to this, that simply comes with experience and feel, and is not something computers can generally pick up on.
However, when properly explained, it then becomes clear that the process is both reasonable and logical. That process is based on something that is not subjective at all, in fact it is based on the accumulated data the figure maker can refer to.
The figure maker uses a base or 'Beyer par' to anchor his numbers. This par is the result of a large enough sample size of races for a particular track, to justify.
There are two separate but key factors involved with the consistency of speed figures.
The first is that the figure must reflect how fast the race was run in relationship to what the variant of the track was that day.
A better way to understand the concept of how a subtle change in track conditions can change the variant for the figure is to look at an example from Preakness weekend at Pimlico.
That friday card featured a host of stakes races capped off by the feature race; The Pimlico Special. This race is a graded stakes race for older horses run at the same distance as the Preakness. As a general rule the older horses in The Special should run a faster race than the three-year-olds in the Preakness. However on this Friday the track was bogged down most of the day with rain, by the time they ran the feature, the sun was out and the track began to dry. This usually results in the track becoming very heavy and slow. In fact the final times of both races reflected that, the Pimlico Special was run in 1:55 4/5ths where the Preakness was a full second faster at 1:54 and 4/5ths.
As predicted this was correct, the speed figure for the winner on that Friday (Revolutionary) was a 106, and the figure for California Chrome was a 105. Now here is where the science of speed figures comes in handy, the Preakness was run in a faster time. Despite the slower time run in the Pimilco Special, when the variant was calculated to the raw figure, the stronger race was the predictably run by four-year-old Revolutionary.
The second factor is that the figure must remain consistent even when comparing various distances or tracks. This is what gives the figures their relevance. The speed figure is essentially adjusting for as many random variants involved in order of obtaining a ‘clean’ number. If there are only four dirt races that day and only one race at 1 mile and 1/8th, how could they get a fair number? Well as Andy Beyer found out from his roommate at Harvard, Sheldon Kovitz, who happened to be a math major. Sheldon Kovitz calculated an way to measure distance variations accurately so Andy could create a parallel time chart.
This premise is what gives speed figures their appeal. The idea is that as distances increase, the value of a length decreases.
Or as stated in Andy Beyer’s ‘Picking Winners’
“A fraction of a second is more significant in a shorter race than a longer race. That led to the creation of accurate parallel time charts that are able to relate times at different distances.”
In the Preakness California Chrome won in nearly similar fashion to the Derby; although he won going away there was a horse bearing down on him, yet there was a decent separation between the second horse and the rest of the field. Yet despite the similarity it was obvious Chrome ran a much stronger race in the Preakness than the Derby. Why is that, well first of all, the distance between the second horse in the Derby (Commanding Curve and the rest was only 1 1/4th lengths, while in the Preakness Ride on Curlin was a solid 6 ½ in front of the rest of the field. Now take into account that the Derby was a 16th of a mile further, which would dictate that the longer the distance of the race, the greater the distance of separation between finishers would be assumed.
To put it in a way that Einstein would make sound sexy, as the distance of the race increases, time and distance of the outcome decreases.