IT'S AN OLD saying that a person who knows how to beat the horse races consistently isn't going to write down his theories in a book and distribute it to the world at large. A good system theoretically is worth more to its originator than book revenues could ever be, so a horseplayer with one worthy of publication should prefer to keep his theories in himself.

But Andrew Beyer is not your normal judge of horseflesh. He is a Harvard man, or at least almost was until he "discovered that the most demanding form of intellectual endeavor exists at the racetrack." The tale of why he failed to get his Harvard degree is a compelling one.

"When I went to college my resolution to become a diligent scholar was undermined by the discovery that four racetracks lay within easy commuting distance of Harvard Square. I managed to dabble at both academic and equine pursuits for the better part of four years, until an irreconcilable conflict arose two weeks before graduations."

It is difficult for any confirmed horseplayer in college to keep his mind on studies at any given moment. But when final exams roll around the task becomes doubly tough:

"My final examination in Chaucer was scheduled on the same day that Kauai King would be trying to win the Triple Crown at Belmont Park. I knew nothing about the Canterbury Tales but I did know something about Amberoid in the Belmont Stakes. So I went to the track."

For those who don't happen to remember, the 1966 Belmont ended in a tremendous upset with the beloved Kauai King reaching the finish line a beaten animal. The winner was none other than Amberoid who paid $13 and had the solid backing of at least one Harvard student who was in the crowd that day. So Andrew Beyer is either very lucky or a psychic. In any case, he made the right choice in leaving Harvard.

Byer's Picking winners is intended for both the novice in search of the basics of horseracing and for the "dead-serious horseplayer wanting to learn how to win consistently." The distinction is more important than that, however. A novice will find the book mildly interesting and occasionally amusing: the hard-core handicapper may memorize its contents.

The chapters which separate the men from the boys, the women from the girls and the racing addicts from the occasional players deal with Beyer's Speed Handicapping formulas, a dense set of figures derived from thoroughbred racing's bible, the Daily Racing Form.

The Racing Form, which may be purchased at newsstands everywhere for the entirely reasonable price of one dollar, is an amazing compilation of the past-performance records of every horse running in races at various tracks on any given day. Horseplayers live and die by the immense amounts of information packed into it, basing selections and theories on its vast resources. Not surprisingly, It is the life blood of Beyer's formula's as well.

A LONG-STANDING dream of handicappers everywhere is to derive a system by which a numerical value can be assigned to a horse's most recent outings--a speed rating by which it can easily be compared to the competition in any race. Thousands of these systems have come and gone, each one failing to take into account any number of abstract inputs which make horseracing what it is: extremely unpredictable.

Central to all systems, however, is the supposed answer to one basic question: how fast can a given horse go? And it is here that Beyer provides some superior insight into how a handicapper can judge past performances and arrive at reasonably accurate speed figures. the key is his explanation of why the speed rating track variant found in the Racing Form is based on faulty correlations and is therefore useless. Beyer circumvents the erroneous derivations of the Form by computing his own speed figures and exhorts the reader to do the same. The process, however, is far from an easy one.

"The procedure is the one phase of making figures that is excruciatingly boring. When I am tackling art unfamiliar track. I closet myself in my handicapping room with a minicalculator, a set of charts of all the races run at the track during the previous year, several big sheets of poster paper. A few sharp pencils, a bottle of Jack Daniel's and a belief that the ensuing hours of drudgery will eventually pay off."

Horseracing and handicapping, then, are both serious businesses. Picking Winners emphasizes a rational intelligent approach to the sport which presupposes that there are indeed some "underlying truths" about horseracing which make it something other than a chance operation.

In the nine chapters of the book not dealing directly with the strict Speed Handicapping theorem. Beyer strings together an entertaining series of short stories about his experiences at the track. That's not so unusual--all horse players love to tell stories of close finishes and amazing upsets. What makes Beyer's yarns different is that they each contain a moral, an underlying piece of horseracing truth. The sum of these truths provides even the novice with a sound basis for winning at the races:

"When a three-year-old is asked to carry the top weight in a race against older horses, he will not win... Of all the characteristics of thoroughbreds, the one that seems to be carried most often from generation to generation is the ability to run well on grass... A handicapper should pay little attention to the sex of horses.