PHILADELPHIA - When Western Playboy won the 1989 Pennsylvania Derby by 17 lengths and ran the 1 1/8 miles in 1:47.60, the Beyer Speed Figure computed to something like a 135. Going into that year's now legendary Breeders' Cup Classic, featuring the showdown between Sunday Silence and Easy Goer, Western Playboy was going to have the top Beyer. But something simply was wrong with that. The time and the subsequent figure made no sense. It simply could not be. One computer-generated service had Western Playboy with the top figure. The Beyer Fig team tried to make sense of it and "projected" the figure to something that seemed more in touch with reality. Easy Goer regularly ran in the 120 range. Sunday Silence did that only when it was necessary for him to win. My memory is that Western Playboy's projected figure was something like 110. Did we know that was right? No. It just seemed to make the most sense. As Sunday Silence and Easy Goer staged their epic stretch run, Western Playboy, sent off at 16-1, was passing Mi Selecto to finish seventh, almost 15 lengths behind the top two. One of the strengths of the Beyers is that there is a human being who examines every one of them. Some days are perfectly straightforward. Every race falls into line and there are no questions. Then, there was June 17 at Belmont Park. The day was what we call 13 Beyer points fast or -13. To get the proper figure for each race, the figure-maker had to subtract 13 points from the raw number (which equates to the actual time of each race). Each race seemed to fit nicely into line - except the third. It was a race for 2-year-old New York-bred maidens. The time (57.23 seconds for five furlongs) equated to a 110 raw number on the scale. After subtracting the 13 points, the actual figure was 97. First-time starter Classic Pack had defeated fellow firster Mr. Sam I Am by a head, so both horses earned a 97. Was it possible for two New York-bred maidens to get such a big figure? Yes. Was it likely that these two horses could earn a figure that would make them among the fastest 2-year-olds in the country? Not really, but you could not truly know because there was no context. In the Western Playboy example, he had a context. These horses did not. So Mark Hopkins, who does the New York figures, entered the 97 into the database and put a question mark next to the figure. Absent any compelling evidence, Hopkins went with the data. "When it's not certain, we try to put a number down that has the best chance of being right," Hopkins said. When Mr. Sam I Am ran back in a statebred maiden race on Aug. 1 at Saratoga, his 97 towered over the field. He was an odds-on favorite. He dueled for the lead and finished fifth, earning a 47. Hopkins adjusted the original figure downward to an 83. When Classic Pack appeared in an open allowance at the Spa on Aug. 7, he went off at 22-1 and finished last, getting a 62 Beyer. Clearly, the first number turned out to be incorrect, but there simply was no way to know that at the time. Perhaps, the two dueling maidens knocked each other out. Thus, their subsequently poor races were a result of that. Perhaps the timer malfunctioned. Perhaps there is simply no explanation. One of the strengths of the figures is that a human being puts a red flag next to a questionable number and, as soon as a horse from the race in question runs back, that number is reviewed. When the original figure proves to be wrong, it is changed. It does not happen often, but it does happen. Consider that thousands and thousands of figures are put into the database every year. The percentage of figures that need to be changed is tiny, but figure-making is art based on mathematics. It is very rare that a figure has to be changed by 14 points. Usually, the change is a few points, up or down. Most of the time, the math is straightforward. Then, there are races like the third at Belmont on June 17. When Lost in the Fog ran out of the TV picture in his second lifetime start, the day after Christmas at Turf Paradise, the data strongly indicated the colt had earned a 109 Beyer. Was it possible that the fastest 2-year-old in America was a son of Lost Soldier training in northern California? Sure. Was it likely? Not really. But the math was the math. That was the figure. It was entered into the database. Days later, I wrote a column extolling the virtues of Lost in the Fog, explaining why the Beyer certainly seemed real. So, before even serious racing fans had heard of this colt, Daily Racing Form readers had a hint that a horse of rare speed and talent was emerging onto the scene.