11/12/2015 1:16PM

Beyer: Change coming to turf Beyer figures


Beyer Speed Figures have undergone relatively few changes during the 23 years that they have been part of Daily Racing Form. Rooted in logic and mathematics, the calculations behind the numbers have stood the test of time.

But we are now making important changes to our calculations for turf races so that these figures will better define the ability of grass runners in relation to dirt runners. With the revised methodology, good stakes and allowance horses typically will earn figures two or three points better than in the past. Many horses at the bottom end of the class ladder will receive lower numbers.

We have reanalyzed and recalculated all grass races since Jan. 1 so that handicappers will be using figures that are consistent with one another. The revised speed figures will be entered into the DRF database on Tuesday night and will appear in the past performances for all races whose entries are taken after that date.

Grass races have always produced low figures relative to dirt. Since they usually are run with a slower early pace, they produce slower final times (when the condition of the racing surface is taken into account). We used to believe that the lower figures for turf runners reflected the realities of U.S. racing.

American Thoroughbreds have historically been bred for dirt and trained to run on dirt, so it was logical that any given class of horses would perform better on dirt than on grass. At Santa Anita from 2005 to 2015, to cite one example, the average winning figure in a bottom-level allowance race (N1X in the past performances) was 93 on dirt and 90 on grass.

As every racing fan knows, turf has become a much more important part of the U.S. sport in recent years. Races on grass regularly draw fields that are larger and more competitive than those on dirt. The higher-class events are filled with runners trained by top grass-oriented horsemen and bred by proven sires of grass runners.

Most classes of grass races now appear comparable in quality to their counterparts on dirt, and we believe our figures ought to reflect this fact. For example, the N1X allowance races at Santa Anita ought to average around 93 on both surfaces.

One exception to this parity between dirt and turf runners might be top-level stakes company. The richest and most prestigious American stakes races are on dirt, and Grade 1 dirt horses in the U.S. remain superior to the domestic Grade 1 turf runners.

Raising figures for horses such as the allowance runners at Santa Anita was not a simple matter of adding points to higher-class races. This was a five-month project during which my partner, Mark Hopkins, and I, aided by our colleague Randy Moss, examined every component of our turf calculations. Jerry Nicholson of The Jockey Club Technology Services and consultant Paul Carter generated volumes of data to aid our analysis.

Like almost all systems of speed figures, ours are based on a “parallel time chart,” a table that assigns a numerical value to every final time at every distance. That number is adjusted by a track variant – which measures the inherent speed of the racing surface – to produce the Beyer Speed Figure. We developed these calculations for dirt races, but they weren’t a perfect fit for turf, producing figures that were too low for the fastest horses while being too high for low-level runners.

In the course of our study, we discovered the technical reasons for these aberrations. We corrected them by modifying our parallel time chart, assigning a slightly higher point value to each fraction of a second. The changes brought our turf figures closer to those for dirt at both ends of the class ladder.

We made revisions to other aspects of the figure-making process. In determining figures for the losers in any race, we have translated beaten lengths into Beyer points by using a formula that hasn’t changed since “Picking Winners” was published in 1975. With previously unavailable data now at our disposal, we have revised the formula to make adjustments that were demonstrably more accurate. A horse beaten by 10 lengths in a route race will earn a figure approximately three points higher than in the past. The new formula also will be applied to dirt races.

No matter what methodology any figure-maker employs, speed figures for turf races will always be more problematic than those for dirt. Tracks’ use of varying rail settings and varying run-ups to the official start of races regularly cause difficulties for handicappers.

However, the most confounding factor in grass races is pace. When the early pace of a race is extremely slow – a common phenomenon – even classy horses might not be able to accelerate enough in the stretch to produce a fast final time. In such a situation, a figure-maker might have to assign the race a number based on the horses’ typical past efforts instead of taking at face value an absurdly slow time. Making figures on grass will never be as precise as it is for dirt.

Nevertheless, as we recalculated all the turf figures in 2015, we were pleased with the overall results. From high-level allowance company to bottom-level claimers, figures on the two surfaces will now be comparable. At Belmont Park, N2X allowance races averaged 95 on dirt and 95 on grass. Maiden $35,000 claimers averaged 74 on dirt and 74 on grass.

Stakes horses on turf no longer look like poor relations of their dirt counterparts. Aside from the effort by American Pharoah in the Breeders’ Cup Classic, the best figures on dirt this year were the 114s earned by Beholder and Liam’s Map. The top performance on turf was the Saratoga victory by the European star Flintshire – also a 114. U.S.-based turf runners produced eight figures in the 108-to-111 range. They’re still not as good as the best dirt runners, but they are closing the gap.

Of course, turf figures cannot predict how a horse will perform on dirt; few Thoroughbreds run equally well on both surfaces. But the revised figures will define the relative talent of dirt and turf runners, and we believe they will be a much stronger handicapping tool as well.