The biases that track surfaces impose on racehorses are important. They play a fundamental and sometimes dominant role in the handicapper�s world. Unfortunately, the specific track biases actually in play in a given situation are usually unknown to the handicappers playing there.

If every race in America was run on the same artificial surface at the same distance in an indoor racetrack, there would be no need for Track/Surface/Distance (TSD) adjustments. Every performance in a horse�s record would be perfectly comparable. Handicapping would be simple, but value would disappear faster than a Kuwaiti's gratitude. As it is, each past performance in a horse�s record comes off a different surface, and the speed of those surfaces vary widely. If you look at the raw running times of each pp, even those run at the same distance, you�ll be fooled time and again. 

If you look at most speed figures, you�ll be no better off. As James Quinn says in Figure Handicapping, "Track-to-track comparisons at the lower levels of racing will be generally reliable, but in higher-priced claiming and in the majority of non-claiming races, the same comparisons become trickier and less dependable." So, you�re left with a lot of less-than-perfect ways of dealing with TSD adjustments.

If you�re a computer handicapper, one of the best ways to deal with Track/Surface/Distance adjustments is to use a program that projects the times for a given paceline so it jibes with today�s TSD. Projected times are usually done by comparing the pars of today�s TSD with the pars of the paceline�s TSD. For instance, if the final time par for the paceline at today�s class level was two seconds slower than today�s final time par, the horse�s time is adjusted faster by two seconds to compensate for the paceline�s slower surface. This works pretty well, given that the adjustments are accurate. If the pars are out-of-date or poorly done, that�s another thing. At any rate, it�s not practical to project times with a paper and pencil � it really requires a computer program.

Now, I know a thing or two about pars. I spent 10 years creating pars for every track in North America. I�m still known to many as "the pars guy." The advertising copy that used to say that those pars took 400 hours per year to create was not lying � it did. And that was with the help of a database program. So I spent about 4000 hours of my life creating par charts for six furlongs at Arapahoe Park and the like. If I had been wise enough to take up the cello instead, I could be playing Carnegie Hall by now. Every April when I was cramming to finish up the annual pars, my eyes would get that red and squinty look favored by H&R Block tax specialists. In fact, I was probably often mistaken for a tax specialist around that time of year. The difference is that the tax code is far more interesting than pars. It would get so bad that I would actually dream fractional race times at night: "22.2, 22.3, 22.4" and so on.  In short, if I never see another par for the duration of my existence on our planet, I will be mightily pleased.

It turns out that I might not need to see another par. Contrary to my belief at that time, I had it proved to me several years ago by a customer that using track models (in other words, a running tabulation of the changing track times) is a superior way to project times. Pars still have their uses, such as providing a baseline to judge each horse against, the creation of speed figures, or lining your birdcage. But for projected times, models are the way to go.

In my current handicapping program, The Capper, I use what I call TSD adjustments instead of pars. Essentially, the claiming times for a given TSD are used to create a claiming times hierarchy. Then the $10,000 claiming times from this hierarchy for the first, second and final call are compared to a universal standard time for each call. The result is three adjustments, one for each call. These TSD adjustments are used, along with other modifications, to create the projected times that apply to today�s Track/Surface/Distance.

What if you don�t use a computer handicapping program that can accurately project times? Another option would be to check out the Cramer speed figures. (See www.rspos.com for more information). From what I understand about the way they are calculated, they use a unique database-oriented method of calculating TSD adjustments, based on comparing horses who have actually run at each Track/Surface/Distance. So it�s possible that these speed figures are superior to Beyers and the like regarding TSD adjustments.

How do you handle Track/Surface/Distance changes if you handicap with nothing but the Form? Try these suggestions on for size:

� Treat speed figures earned at the same circuit as today�s race with more confidence. Use some old-fashioned class analysis when comparing a horse from out of town.

� Give little credence to turf speed ratings. Especially, don�t use a turf speed rating when handicapping a dirt race. Interestingly, the converse doesn�t hold: dirt speed ratings can be very good predictors in turf races, assuming the horse doesn�t hate the grass.

� Remember that speed figures from one distance can be mathematically accurate but meaningless when applied to another distance, for instance, if the horse loves to sprint and hates to route. Two-year-old sprint speed figures are not the best predictors when the young-uns stretch out in the fall.