Beyer Speed Figures

Since making their published debut in 1992, Beyer Speed Figures have become the industry standard for evaluating Thoroughbred performance.

Ask Beyer Q&A Archive - Note: Here is a sample of some past questions answered by Andrew Beyer. Submit your question to askbeyer23@drf.com

Not long after past performances for Saratoga’s opening day appeared on the DRF website, I received a call from a handicapper who declared: “You have a bad mistake on the figure for Queen Arella in the Schuylerville Stakes.”

I cringed. We hate making errors anywhere, let alone on one of the most high-profile racing days of the year. But my caller’s evidence appeared correct. Queen Arella had won her debut at Gulfstream Park impressively, by four lengths, but her figure was a paltry 44. My caller had looked at the Beyer Speed Figure data in DRF Formulator and saw that the 2-3-4-5 finishers behind Queen Arella had each improved by 20 points or more in her subsequent start. This is the type of evidence that will often lead us to conclude that a figure is erroneous and change it retroactively.

However, I looked further at the Gulfstream data and found this: On May 29, when Queen Arella ran, there was another five-furlong race on the Gulfstream card, for $12,000 claimers. The winner, Combination, was clocked in 57.16 seconds and earned a figure of 84. Two races later, Queen Arella covered the same distance in 59.54.

The 2.38 second difference between the two races equals 40 Beyer points -- thus the figure of 44 for the latter race. The speed figure was apparently correct. But why has everybody in the field improved so dramatically?

I make the speed figures for Southern California, and I was perplexed all spring by the low numbers for 2-year-olds running 4 ½ furlongs at Santa Anita. Our calculations for such races had always made sense in the past -- the average for open maiden special weight 2-year-old races from 2016 to 2019 was 66. But this year most juveniles ran significantly slower.

On May 25, a colt named Hendavid won an open MSW race with a stunningly low figure of 36. But as in the case of the Queen Arella race, all of the horses in the field improved when they ran again. The 2-3-4-5 finishers increased their figures by an average of 20 points.

What is happening? COVID-19 is happening. The pandemic has, of course, disrupted racing across the country, causing extended cancellations as well as uncertainty about future racing schedules. Trainers of 2-year-olds may have felt that there was no reason to subject their youngsters to hard training with the future in doubt. If 2-year-olds made their racing debut with less preparation that they would have in a normal year, they might reasonably earn poor figures and then improve significantly with the benefit of racing experience. That’s my theory. In any case, handicappers should be a little cautious in evaluating 2-year-olds who have begun their careers with an abnormally low figure, and we should be alert to the possibility of sudden improvements.

Postscript: Queen Arella reportedly injured a foot while running fifth in the Schuylerville Stakes,earned a Beyer Speed Figure of 42.

Q. A horse runs in an optional-claiming race and gets a figure of 72. Next, the trainer takes a shot in a stakes to help the track fill the card. The winner runs a 95, and our horse finishes last but gets a new high Beyer of 79.

I call that the “suck along” number. Is this in the character of horses to try to keep up with much better runners?

–Gordon Osborne, Louisburg, N.C.

A. Were you a harness bettor at some time in your life? What you describe is a phenomenon well known in harness racing.

Because wind resistance is such a big factor in the sport, trotters and pacers regularly line up in single file, and an overmatched horse on the tail of this line, shielded from wind, may record a fast final time that he could not duplicate in a competitive race.

In my opinion, such scenarios do not arise in Thoroughbred racing. Horses do not get any tactical or psychological advantage by racing against superior rivals. Horses will often earn worse-than-usual figures if they are overmatched. If a horse possesses a modicum of speed and tries to chase high-class speedsters, he may tire so badly that he runs a disastrously bad race. And if a horse is trailing the field in the stretch, his jockey may not bother to urge him as he would if the horse were battling for the lead.

To return to your example: If a horse runs a figure of 79 finishing last in a stakes, and then drops into a more realistic spot where 79 is the top figure, I would not hesitate to bet him.

Q. Just curious about figures at Belmont on June 18. Its All Relevant earned a 99 Beyer Speed Figure going 1 1/16 miles in 1:39.87. So Darn Hot earned a 75 Beyer figure going 1 1/16 miles in 1:41.17. That’s a difference of 1.3 seconds. On the surface it seems that their figures should be about 10 to 12 points apart. Why is there a 24-point difference?

– George Weaver

George, your math is good. The two final times equate to a 13-point difference in our figures. And we understand that you would feel your winning filly didn’t get the credit she deserved.

The explanation derives from the fact that Belmont can be a very changeable track. Sometimes it will speed up significantly for a single race. We can never be certain when this has happened, but we can usually deduce it if a fast final time would produce implausible figures for most of the horses in the field.

A case in point was Tap It to Win’s victory in a June 4 allowance race. Our normal calculations would have given the race a figure of 107, with the top six finishers all improving significantly upon their career-best efforts, Tap It to Win would have looked like a standout in the Belmont Stakes if we had accepted the figure. We arbitrarily assigned the race a 97, which better reflected the talent of the field. Tap It to Win’s fifth-place finish apparently confirmed that we were right to lower the figure.

The first race on June 18 was a similar situation. If we made our calculation based on the track variant that applied on the rest of the card, and which gave Its All Relevant a 99, So Darn Hot would have earned a winning Beyer Speed Figures of 86.

The filly had earned a 66 in her one previous start on the dirt. Could she have run so much better? Sure. She had been a $600,000 purchase and she would be expected to improve with experience.

If the winner ran an 86, runner-up Thankful would get a figure of 77. She had shown nothing in her debut on the turf (a fig of 60), but she had a strong pedigree and could be more effective on dirt.

Third-place Sky Queen would receive a figure of 76. She had raced twice at Aqueduct, showed speed in both starts and faded in both, earning figures of 50 and 54. Could she now improve more than 20 points? Hmm.

This was a judgment call, and not an easy one, but we thought it was unlikely that all three fillies had improved so much at the same. We assigned the race a figure of 75, which better reflects their previous form while allowing for some improvement. We might be wrong. In such cases, we monitor the subsequent performances of the horses in the field and review our initial judgment. We do not hesitate to revise our figures. If So Darn Hot comes back to run in the mid-80s, we would retroactively give her June 18 performance an upgrade.

In some races, the run-up distances can be rather large – especially at Gulfstream. When making your figures, do you add the run-up distance to the distance of the race?
–Ray Howard, London, Ontario, Canada

In dirt races, the effect of a long run-up is factored into the Beyer Speed Figures. Some tracks employ a long run-up so that races don’t start too close to a turn.  This is the case for six-furlong races at Churchill Downs, and so the starting gate is positioned about 180 feet behind the pole where the timing of the race begins. (At other distances, the run-up is 34 feet.) The Churchill six-furlong races usually produce fast fractional times, and fast final times as well. We have a mountain of data showing that the six-furlong races are five Beyer points faster, in relation to 6 1/2  and seven furlongs, than they would be on a track with a normal configuration. These five points are built into the calculation of our speed figures.

I wish that I could offer such a clear, confident explanation of run-ups in turf races. The main difficulty in making figures on turf is the effect of pace on final times. When horses crawl the first half-mile in 50 seconds in a one-mile race over a firm course, it’s usually impossible to compare the final time to a race with a sensible pace – even if both are run at the same distance with the same run-up. We have to use our judgment in making figures under such circumstances, and adding the factor of a long run-up makes judgment calls more difficult.

As you observed, the conditions of turf races at Gulfstream can be daunting. Earlier this winter, for a one-mile turf race, the temporary rails were positioned 120 feet into the course, and to accommodate the changed circumference the run-up to the start was 282 feet. There is no historical data on which to base speed figures for such a setup. Moreover, at many tracks, turf races with long run-ups and wide rail settings produce final times that I suspect are inaccurate. I wish U.S. tracks would employ the same run-up distance used in much of the world: zero.

Q. As someone who worked as a handicapper (providing what I think you would call “class ratings”) for the British Racing Authority, I am interested to know why you appear not to take into account weight carried by horses when producing your speed figures.
–Stephen Hindle, Miami Fla.

I developed speed figures in the 1970s when I realized that time was the most important single factor in the game, and that I needed precise calculations to compare horses’ times over different distances and different tracks. I never aspired to create a figure that was an all-encompassing definition of a horse’s ability (what you might call a class rating). I felt that if I had an accurate speed figure, I could then use my own handicapping to deal with other key factors such as trips, pace, track bias, trainers, etc.

I never paid much attention to weight, but I got a fast education in the importance of weight in the early 1990s when I spent a few months playing the races in Australia. Many of the races there were handicaps in the truest sense of the word; it was not unusual to see the top-weighted horse carrying 140 pounds in a run-of-the mill race. After considerable study, I concluded that one pound equaled four-tenths of a point on the Beyer scale. Thus, if a horse was assigned to carry 20 pounds more than he did in his prior start, I would subtract 8 points from his previous figure. If you’d like to incorporate weight while handicapping with the Beyer Speed Figures, I’d recommend that adjustment.

Q. Beyer Associates must on occasion “project” a figure, given the pace dynamics of a race. Considering that a projected figure carries much less weight than one generated by data, is there a way to distinguish the differences in the past performances, such as italic type or a font that’s not in boldface?
–Tony Bleill, Champaign, Ill.

Your question is a good one, but what you suggest is something we would never do.

There are various circumstances when the final time of a race is so fast or so slow that a true speed figure would make no sense and would only mislead our readers. In such cases we project a figure – we assign the race a number that does make sense and properly reflects the ability of the horses in the field.

Aberrant times can arise when the inherent speed of a racing surface changes during the course of a card, for no apparent reason. (Track superintendents can speed up a track significantly by watering it more heavily than usual.) They can arise when the track’s timing system misfires. They can occur when the early pace of a race is so slow that the horses cannot possibly accelerate to produce a meaningful final time. This is a common scenario in turf races.

What should a figure-maker do with a race such as the 2020 Sunshine Millions Filly and Mare Turf? Starship Jubilee, a Grade 1 stakes winner, captured the event at Gulfstream Park on Jan. 18, running 1 1/16 miles in 1:42.57. One other race was run on the card under the identical conditions (with the rail setting at zero). It was a $16,000 claimer, and the winning time was 1:41.66. In the Beyer system of numbers, the claiming race was 11 points faster than the stakes. The claimers figured to run a figure of 83 (at best), which would make Starship Jubilee’s figure a 72. Not only would the mare be running 30 points below her best form, but every other member of the field would be regressing 15 to 20 points. Obviously, a winning figure of 72 would be inconceivable.

The explanation for this anomaly was obvious. Starship Jubilee was allowed to set a crawling pace, running the first six furlongs in 1:14.66, three full seconds slower than the fraction for the claimers. Mark Hopkins, my partner who makes the Gulfstream figures, didn’t hesitate to project the figure as a 93. This was below the winner’s top efforts, but it gave figures to the seven fillies behind that were in line with their previous form.

Should we have put the figure for the race in italics because it “carries less weight than one generated by data?” The answer is no. Our aim is to publish numbers that accurately define horses’ ability and will therefore be useful to our readers. A projected figure that accomplishes these aims does not deserve to be stigmatized. And a preposterous figure is no less preposterous if it is “generated by data.”

Will Beyer Speed Figures in May 2020 be helpful for handicapping a September 2020 Kentucky Derby?
– Hugh Ernstberger

Over a four-month period, a lot can happen to 3-year-old racehorses. They can improve dramatically, they can regress, they can be compromised by injury, so it would be unrealistic to make a forecast for the rescheduled Derby on the basis of speed figures earned in the spring. Nevertheless, the early season performances by Charlatan could well point out the 2020 Derby winner. He recorded Beyer Speed Figures of 105 and 106 in his first two career starts at Santa Anita – better than any member of his generation. A 106 would have been good enough to win any Kentucky Derby since Big Brown’s victory in 2008.

But the historical norms of the Kentucky Derby may not be relevant on September 5, 2020 at Churchill Downs. Running 1 1/4 miles on the first Saturday in May has always been a test of horses’ fitness as well as their raw talent. In an era when trainers campaign their horses sparingly, colts now come into the Derby with preparation that would have been deemed grossly inadequate for most of the race’s history. Over the last 15 years, the winning speed figures in the Derby have been mediocre, because the lightly raced entrants don’t have the seasoning to deliver great performances.

This evidence does not necessarily mean that U.S. Thoroughbreds are in decline. In 2016 Arrogate was a late bloomer who missed the Derby but earned a speed figure of 122 winning the Travers Stakes in August. American Pharoah’s Triple Crown victories in 2015 were unexceptional, from the speed-figure standpoint, but he recorded a 120 in the Breeders’ Cup Classic in October. Thoroughbreds usually don’t reach peak form until the fall of their 3-year-old season – or later. If the winner of the 2020 Derby is not Charlatan, it is apt to be a colt who improves significantly in the second half of the year.

I have been told that you tried to apply your speed figures to Australian racing but it did not work out. I’m stunned at the lack of info available in Hong Kong, Japanese, and Australian races. You’d expect that they’d be more advanced in handicapping than the U.S. Do the players there look at racing as pure luck? Why did your attempt in Australia fail?
–Jason Bergstrom, San Diego

The past performances that U.S. players see for international races probably gave you a false impression of racing in Australia, Hong Kong, and Japan. In each of these countries there are innumerable publications and websites offering a vast amount of racing data. Much of the betting public takes handicapping very seriously.

On visits to the tracks in Hong Kong, I have always been struck by the quietness of the big crowds. Everyone is carrying at least one racing publication; many are wearing earphones to listen to radio commentary on the races. The fans there go about the business of gambling in deadly earnest.

I wrote about my 1991 Australia adventure in “Beyer on Speed.” After developing speed figures for the country, I lived in Sydney for three months and played the horses intensely. The figures enabled me to break even (actually, I won $500 – before expenses) but they certainly didn’t give me an edge. If a horse with a superior figure went off at 10-1, the odds were almost always right and the figs were wrong.

I had difficulty adapting to a country where all the races were run on grass, but the main fact of life in Australia (then and now) is that the wagering competition is so tough. Tote odds usually fell in line with the prices posted by on-course bookmakers, and bookies couldn’t stay in business without being very smart and well informed. I knew one of savviest gamblers in Australia, and even he could only hope to grind out a narrow edge versus the bookmakers. If he rated a horse as a 2-1 chance, and he was able to get 5-2 on his wager, he was ecstatic.

I don’t know if there are any places left on Earth where a studious handicapper can find a significant edge, but I do know that Australia isn’t the place to look.

Q. The Louisiana Derby was run at 1 3/16 miles in 1:56.47, earning a 91 Beyer Speed Figure. The New Orleans Classic, run earlier on the card, was run in 1:50.27, earning a Beyer Speed Figure of 98. This seems very odd as it would equate, approximately, to 1 3/16 miles in 1:56.80, slower than the Derby time. Unless there was some incredible speed-up in the track, it seems to me that the Louisiana Derby Beyer Figure should be much higher than 91. In fact, I would estimate it to be about 102, assuming the track speed was the same for both races. Your thoughts?
–John M. Heft, Aliso, Viejo, Calif.

The speed of the Fair Grounds track was consistent throughout the card of March 21. The Beyer Speed Figure of 98 made sense for the horses in the New Orleans Classic, and you are correct that the final time of the Louisiana Derby, won by Wells Bayou, was several points faster than the Classic.

But if you look at the other races on the card, you’ll see the problems that confronted Randy Moss, who has made our Fair Grounds figures for many years. “If the Louisiana Derby gets a 102,” Randy said, “we must give Bonny’s South’s win in the Fair Grounds Oaks a half-hour earlier a 98.” (None of the top finishers in that field had ever earned a figure as high as 80.) “We’d have to give a maiden race for 3-year-olds a 97.” (The winner Mystic Guide had previously run a 72.)

Randy concluded that the key to understanding the day was the slow pace of the New Orleans Classic. This was a solid field of older stakes horses, but the fractional times for the race – a half-mile in 50.09, six furlongs in 1:14.36 – were slower than any of the five other two-turn dirt races on the card. The winner, By My Standards, should have been capable of running faster than the Louisiana Derby, but after the slow pace he couldn’t do it.

If we dismiss the slow final time of the Classic as an aberration, all of the other races make sense, with Wells Bayou earning a 91 in the Derby, a number that was consistent with the previous best efforts of the 2-3-4 finishers in the field.

But what do we do with the New Orleans Classic? A figure lower than 91 wouldn’t have made sense for By My Standards or the runners behind him. In cases such as this, where the final time has been significantly affected by a slow pace, we “project” a figure – we assign the race a figure that reflects the ability of the horses in the field. We believe that the figures of 98 for the Classic and 91 for the Derby do just that, though we would be happier if the data for the day had been more straightforward.

How did handicappers [of the past] handicap without speed figures? Did they just compare finish times, or what?
– Bradley Bollhagen

If you can find a copy of Toney Betts’s “Across the Board,” you can read an engaging history of gamblers and assorted racetrack characters from the 1920s through the 1950s.

Betts wrote that a handicapper named Colonel William James timed races by himself, hired an engineer to compute wind velocity and calculated his own track variants. He won a fortune at Belmont Park, Betts said, but “lost it on the Floozies of Broadway.” The most famous gamblers of the 1940s were Jules Fink and a syndicate known as the Speed Boys, who made fractional times a centerpiece of their handicapping approach.

Q. I’ve been using your speed figures for at least 20 years with good success. What is the most important figure – the last one, an average of the last three or four, or something else?

– Michael Tucci, Harrisburg, Pa.

A. There are plenty of valid ways to use speed figures, but calculating an average is not one of them because a single aberrant result can create a totally misleading average. It’s much more sensible to look at a horse’s performances individually. If his three most recent figures are 40, 70, 70 and there are no extenuating circumstances to explain the bad last race, a handicapper might conclude that the 40 represents the horse’s current ability. If the horse had an excuse in the last race, the handicapper might reasonably decide that he’s ready to run a 70 again. However, an average of 60 for the three numbers is in no way meaningful.

I begin to study a race by looking at each horse’s most recent figure, identifying the top last-race figure in the field. I can usually disregard rivals who have proved that they can’t compete with this top figure. In some cases, handicappers should look well past a horse’s most recent start if he has not been racing under the right conditions. If a horse has been competing recently on grass, on sloppy tracks, or in route races and his best form is dirt/fast tracks/sprints, a bettor might focus on a figure earned weeks or months ago.

In any case, I try to look at a figure in the context of the way a horse earned it – i.e., what kind of trip he had. If a horse receives a figure of 80 after a hard head-and-head duel for the lead, I’ll think: “This effort was better than an 80.” If he earned an 80 after getting an unchallenged lead and setting a slow pace, I may conclude that he has little chance of duplicating that figure. I am a devout believer in our speed figures, but handicapping judgment must sometimes override them.

–Andrew Beyer

Q. I always wondered why in Daily Racing Form there are no Beyer Figures for foreign horses. It has always made me wonder if you and your staff create figures for foreign horses [for your own use] but don’t offer them to the public.

–Lenny Ficarelle, Massapequa Park, N.Y.

A. I would like to say that my colleagues and I are quietly making a fortune with our Longchamp figs, but this is not the case. The Beyer Speed Figures were designed for American racing, with its relatively uniform oval dirt tracks, and I would not know how to deal with the irregular, undulating turf courses of Europe.

Nevertheless, I have always been intrigued by the idea of making figures abroad, and over the years I have tried them in Australia (with so-so results), Hong Kong (poor results), and Sweden (terrible results.) In 2008 I spent month playing the horses in Argentina, where most of the sport consists of speed-oriented dirt racing that isn’t so alien to an American handicapper. I devoted a full year to developing figures for the country’s three principal tracks—a massive task. It is challenging enough to create a set of figures for a new U.S. track, and in Argentina I was starting from scratch, using information in Spanish (which I don’t read.) But by the time I went to Buenos Aires, I had almost as much confidence in my figures for Palermo as I do for those at Belmont Park.

The result: I didn’t make a peso. The takeout in Argentina was ridiculously high. The betting pools were too small to accommodate serious wagers. Handicapping data was inadequate. There was no betting from home computers. And there was no practical way to turn Argentine speed figures into a commercial product. My year of intense preparing had been a total waste. Even so, I can’t fully suppress the thought that either South Korea or Dubai might be fertile ground for international Beyer Speed Figures.

–Andrew Beyer

Q. In your opinion, why have the Beyer Figures for 3-year-olds headed to the Derby dropped so far in the past decade to 15 years?  

–Tony Mastropietro, Levittown, Pa.

A. The decline in speed figures earned by 3-year-olds in the spring has certainly been dramatic. The last five runnings of the Kentucky Derby produced winning Beyer Speed Figures of 103, 102, 103, 105 and 97. Twenty years earlier, in the period from 1994 to 1998, the winning figures ranged from a low of 107 to a high of 115 (Silver Charm.)

Our figures have not changed during this time. And there is little reason to believe that U.S. racehorses are in decline. In the last five years, horses such as Arrogate (122), American Pharoah (120), Shared Belief (115) and Bayern (113) have delivered outstanding performances in the latter part of their 3-year-old campaigns.

So why have speed figures in the Triple Crown series and the Kentucky Derby prep races been so low? I can offer a theory. For most of the Derby’s history, fitness and racing experience were essential for success on the first Saturday in May. Horses almost always had a solid foundation of experience as 2-year-olds. (Secretariat raced nine times at 2. Carry Back made an astonishing 21 starts as a juvenile before he went on to win the 1961 Derby.) Three-year-olds came into the Derby ultra-fit. Whirlaway ran in two major stakes in the nine days before the 1941 Derby—a common practice at the time.

In the modern era, of course, training styles have changed. Talented young horses are raced sparingly. They have short 2-year-old campaigns and may race only two or three times as 3-year-olds before the Derby. Their final prep race may be scheduled four or five weeks before the Derby (an unthinkable idea for old-time trainers). Justify was unraced at 2 and made only three starts prior to his sweep of the 2018 Triple Crown. He was a brilliant talent, but with so little experience he couldn’t have been expected to run as fast as great 3-year-olds of the past had done in the spring classics. That’s the reason his figures in the Triple Crown were so low—103, 97, 101.

–Andrew Beyer