The headline in the Daily Racing Form read: "Cal Board to Investigate Longshot Maiden Winner."

After Cielo Canosa scored an improbable 39-to-1 victory at Del Mar, under circumstances suggesting he was the vehicle for an audacious betting coup, the California Horse Racing Board undertook a probe of the race. One might assume that horseplayers would be grateful that the sport's officials were looking out for the public interest. But the opinion of most fans here is quite different. They view the alleged coup with grudging admiration. And they believe the CHRB is looking in the wrong direction if it wants to protect the sport's honesty.

I fully share these sentiments. I'd give my Eclipse Award vote for 1999's outstanding trainer to F.A. "Red" Lowery, who won his only race of the year with Cielo Canosa. If there were an Eclipse Award for wise guy of the year, he would win it by acclamation.

Almost every trainer dreams of putting a horse into a spot where he can't lose, and hiding his form so well that nobody else knows about him. There's nothing illegal or immoral about trying to do so, but it's extremely difficult because racetracks are filled with sharpies watching the moves of other sharpies.

When Cielo Canosa appeared in a maiden claiming race at Del Mar, his record showed one previous start at a track so obscure that most horseplayers haven't heard of it: Grants Pass, in Oregon. There the 4-year-old trailed the field all the way in a maiden race, losing by nearly 15 lengths. Unable to beat a single horse at the bottom level of the minor leagues, Cielo Canosa showed up at Del Mar. What was he doing here?

The answer was evident moments after the race began. Cielo Canosa exploded from the gate, led all the way and scored by four lengths, paying $80.20 to win. This was no fluke; the colt ran six furlongs in 1:10 flat, earning a Beyer Speed Figure of 93 that would win just about any race at this level. F.A. "Red" Lowery had the goods.

A few days later, the Daily Racing Form reported that the Caliente Race and Sports Book in Mexico--where the wagers don't go into the Del Mar pool--had taken some $4,000 in action on Cielo Canosa. While many horseplayers suspected that the horse received a less-than-energetic ride in Oregon, Lowery offered a different explanation for the form reversal. He said he had been given the horse to train after his defeat and strengthened his bad knees by working him on a high-speed treadmill.

Whatever the explanation, I hope Lowery broke the bank in Mexico. Although I frequently pontificate in print about protecting the betting public, I don't get indignant about horses such as Cielo Canosa. If a horse is "given a race"--a common practice with first-time starters--bettors have a fair chance to figure out what's going on. I have made some of my most memorable scores by spotting a jockey putting a horse under a stranglehold, then betting him the next time he runs. In this case, a bettor might have been intrigued by information in the official result chart of the Grants Pass race: "Cielo Canosa was unprepared at the start, jumped shadows in the stretch and raced wide throughout."

Most horseplayers don't object to maneuverings by wise guys as long as they have a chance to capitalize on them. But fans become incensed when they believe they don't have a fair shot and that insiders control the game. This summer, horseplayers at Del Mar are angry, indeed, about races in which a trainer starts two horses, uncoupled in the wagering, in a small field. Throughout the history of thoroughbred racing, horses trained by the same person have usually run as a single betting unit, because of the obvious possibilities for dishonesty. But in California such horses run uncoupled in the wagering. Because California is plagued by a shortage of horses, and because Bob Baffert operates the dominant stable here, Baffert often starts two prominent contenders in a small field. And it's driving everybody crazy.

On opening day, Baffert had the 6-to-5 favorite, Coldwater Canyon, in a six-horse maiden field, along with a colt, Captain Steve, who had been soundly beaten in his only start. Confounding most handicappers, Captain Steve won, paying $8.80, while the favorite ran poorly. In another maiden race, Baffert's highly regarded 2-year-old, Gibson County, was the 4-to-5 favorite; the trainer's assistant appeared on in-house television and extolled the colt's virtues. Gibson County finished next to last, while the runaway winner, at 11 to 1, was Stormy Jack--the other Baffert entrant in the field. The higher priced of two Baffert horses has won half a dozen times at Del Mar, and many horseplayers suspect a dark conspiracy.

I'm as paranoid as any horseplayer, but I don't believe that Baffert, with millions of dollars of horses in his barn, is spending his time plotting betting coups. But when a trainer finds himself with the two best horses in a field of six, he is in a position to know who is likely to win the race and, indeed, to control the outcome. In a game where everybody likes to cash a bet, any trainer would be expected to take advantage of such an opportunity.

Fans will suspect conspiracy when anyone trains an uncoupled entry. Whenever racetracks have allowed uncoupled entries (as Maryland did in the 1970s) the practice has generated immense suspicion and hostility. The problem here isn't Baffert. The problem is the California rules, and racing officials need to re-examine them.

An occasional coup such as the victory of Cielo Canosa adds a little spice to the game; moreover, it requires imagination, boldness and skill to execute. But it doesn't take much creativity for a trainer to make a score when he has the two best horses in a small field. Allowing uncoupled entries under such conditions is like granting a license to steal.