10/31/2006 1:00AM

Changing a wrong number for all the right reasons


PHILADELPHIA - When you compute thousands of Beyer Speed Figures annually, mistakes will be made. This is not an exact science. It is very much art. When mistakes are made, they can be corrected.

Some figure-making systems are computerized. When computers make mistakes, they go uncorrected.

A human does every Beyer Figure at every racetrack. The final number is the classic "best version of the truth." It is not always the only truth.

This brings us to July 14 at Belmont Park. A 2-year-old daughter of Storm Cat won the fourth race by four lengths, running five furlongs in 57.13 seconds. After looking at that race in relation to the other races on the card, a Beyer of 102 was assigned to Magical Ride. For weeks, Magical Ride's name appeared on page 2 of this paper because she had the best figure by a 2-year-old in 2006.

The variant that day was -9, so the track was a little fast, but not really fast. The Magical Ride race had all first-time starters, so there was no baseline for those horses and no reason to question the number. The rest of the races fell perfectly in line. As horses from the other races began to run back, the variant held up. Horses confirmed their figures, except in the Magical Ride race.

After horses from that race started to run back, there was reason to question the number. Magical Ride came back a month later as the 4-5 favorite in the Adirondack at Saratoga. She got into a speed duel and tired in the stretch to finish third, earning a 66 Beyer. On Sept. 23, she was the 6-5 favorite in the Matron at Belmont. She finished fifth, earning another 66.

So what happened?

Five Star Daydream, second to Magical Ride in that maiden race, got an 89 that day. She won her next two starts with figures of 79. Nobody else in the field came close to confirming Magical Ride's 102.

Something was clearly amiss. The 102 was wrong.

Andrew Beyer studied all the five-furlong races at the meet and found that several of them came back unusually fast.

Perhaps, it was the run-ups, which seemed especially long, thus giving horses what was essentially a head start into the first quarter. Those run-ups may have skewed the five-furlong figures at Belmont.

Beyer and fellow figure-maker Randy Moss studied the run-ups, but the run-ups in 2005 were even longer and the five-furlong figures from that meet made sense.

For whatever reason, the fast times at five furlongs were an aberration. Several of the five furlongs numbers were lowered.

Magical Ride's 102 became a 92.

"We are usually reluctant to change high-profile figures," Beyer said. "And we certainly weren't happy about lowering Magical Ride's 102 after it had been cited in page 1 news stories and listed as the best 2-year-old number of the year. But by the end of September, the evidence was overwhelming that her fast time was an aberration, and we couldn't let the figure of 102 stand."

Any time there is a questionable number, the Beyer figure-makers note that so they can check back when the horses from that race begin to run back. The vast majority of numbers hold up. Some do not. There may be an obvious reason, there may be no reason. In the end, the evidence is the evidence, and corrections are made when necessary.

If you have played the game long enough, you know that the answer sometimes is that there is no answer. You just know that you got it wrong. And when you get it wrong, you have two choices - be stubborn or get it right.

Magical Ride's 92 makes sense. Five Star Daydream got a 79 in that race and in her next start. Cherokee Sheik got a 72 that day. She won her next start and got another 72. The fifth horse, Trippiontherug, got a 53 that day and another 53 in her next start.

So what happened to Magical Ride? Why couldn't she run back to the 92? If we knew the answer to questions like that, this really would be an easy game. Again, sometimes the answer is that there is no answer, other than that is no easy game.