When Lemhi Gold was named the champion older horse in America last season, the Daily Racing Form analyzed his pedigree in its typical inscrutable fashion.

"Lemhi Gold, a true stayer, has a Dosage Profile leaning heavily to the right, the stamina wing," the paper wrote. "It reads: 3-6-7-0-16, giving him a powerful Dosage Index of 0.64 . . . "

Most readers either dismiss or ignore this mumbo-jumbo. But this week the man who created these esoteric terms and numbers came to Los Angeles to lecture on his methods, and the skeptics who expected him to be a crackpot were sorely disappointed. Steven A. Roman's method of analyzing thoroughbred breeding is, at the very least, fascinating. And, far more often than not, it seems to be correct.

Roman, who has a Ph.D. in chemistry and a strong background in biology and mathematics, became interested in breeding when he was raising show horses in California. So he immersed himself in the literature of the field, and was attracted to what is called the "dosage system."

A European breeder, Lt. Col. J.J. Vuillier, laid the groundwork for the theory when he looked back four or five generations into the pedigrees of top racehorses and found that the names of a handful of stallions kept reappearing in them (look at the bloodlines of almost any good modern horse and you'll probably find Bold Ruler or Native Dancer somewhere). Eventually the stallions capable of leaving their mark on a generation are replaced by a new set of dominant stallions. "Vuillier believed that this is the mechanism that defines the breed and its evolution," Roman said.

Vuillier called these dominant stallions "chefs de race" and evaluated horses by counting the number of "chefs" in their pedigree. An Italian breeder, Dr. Franco Varola, took Vuillier's method a step further, dividing the "chefs" into different groups according to their characteristics, so that he could judge whether a horse's pedigree would endow him with exceptional speed or exceptional stamina.

After reading the theories of both these experts, Roman said, "I thought there were gaps in their approaches--and they were gaps that I would be able to fill." For five years he studied thoroughbred pedigrees, trying to find an answer that men have sought since the breed began: How do you breed a horse to get the ideal blend of speed and stamina? Roman not only found an answer, but distilled it into one neat number.

Roman began by using Varola's categories which divided "chefs de race" into five types, according to the natures of the offspring they sired:

Brilliant--Horses with speed and precocity, but no distance-running ability.

Intermediate--Horses with speed and early maturity who sometimes can handle classic distances.

Classic--Horses with a balance of speed and stamina. ("These are the foundation of the breed," Roman said. Stallions like Northern Dancer, Exclusive Native and Nijinsky appear in this category.)

Solid--Horses with less speed who mature late.

Professional--Plodders with no speed and great endurance.

Roman looked at a horse's pedigree for the last four generations and assigned points for each "chef de race" in it. If the horse's sire was a "chef" he counted 16 points. If his grandsire was a "chef," 8 points. A "chef" who appeared in the third generation back was worth 4 points, and one in the fourth generation 2 points.

Now Roman tallied the total number of points for Brilliant "chefs," for Intermediate chefs and so on. In the case of Lemhi Gold there were 3 points for Brilliant stallions, 6 for Intermediate, 7 for Classic, 0 for Solid and 16 for Professional. His dosage diagram was said to be 3-6-7-0-16, with a heavy emphasis on the Professional category. Lemhi Gold was bred to be a plodder, and, indeed, he was a horse who could rarely win at a distance shorter than 1 1/2 miles.

To make these figures more wieldy, Roman distilled them into a single number. He calculated the speed influences in a pedigree by adding the number of points in the Brilliant and Intermediate categories to half the points in the Classic category (which, after all, was supposed to represent the ideal balance of speed and stamina). To measure stamina, he added the Solid and Professional points to half the points in the Classic category.

Then he divided the number of speed points by the number of stamina points to calculate the Dosage Index. The higher the number, the more speed-oriented was a horse's pedigree.

While the skeptics make much of the occasional exceptions to Roman's figures, such as Conquistador Cielo's ability to win at 1 1/2 miles, the numbers usually do reflect horses' capabilities. Lemhi Gold's Dosage Index is 0.64. Perrault and April Run, long-winded runners who were voted the country's champion grass horses last year, had indexes of 1.13 and 1.10. In contrast, the Eclipse Award-winning sprinter Gold Beauty had a Dosage Index of 3.00. The brilliantly fast Landaluce was a 3.67.

With the Dosage Index as his guide, Roman started analyzing the performances of large numbers of horses.

"When you look at the elite horses--champions, classic winners, million dollar earners--only about 10 percent have a dosage index above 4.0. The index of superior horses is usually around 2. This is where I started getting into the soup. I was saying, essentially, that the classic horse is bred a particular way, and that upsets people who have their own breeding program and breeding ideas."

Yet when breeders express their skepticism, Roman confidently counters them with facts. No horse in history with a Dosage Index over 4.0 ever has won the Kentucky Derby; only two ever have won the Belmont Stakes. "With this method," he said, "we can define a horse's character. We have found a way to relate his pedigree to his performance."