Getting up to Speed on BeyersBecause determining who's the fastest isn't the only use
By Dana Byerly, Hello Race Fans Co-founder
Beyer Speed Figures are the most well-known tool when it comes to assessing a horse’s speed, in part because their creator, Andrew Beyer, is one of the most well-known and respected handicappers in America. “Figures are everywhere, so you might as well understand them” was the opening line in Ed DeRosa’s Figuring Out Figures post, and now we’ll help you get a better understanding of Beyer Speed Figures, or BSFs.
Instead of having a single person write a piece on BSFs, I thought it would be more instructive (and fun!) to have a discussion with folks who run the gamut of Beyer/fig usage. I’ve asked Adam Wiener, Ed DeRosa, Jessica Chapel, Chris Rossi and Ernie Munick to share their insights, uses and thoughts on BSFs. Since I don’t use Beyers or figs of any kind, I thought I would be the perfect discussion moderator… let the fun begin!
First off, thanks for joining in me in what I’m sure will be an informative and entertaining chat about the ins and outs, pros and cons, pains and pleasures of Beyer Speed Figures. To level-set our readers please describe your usage and reasoning. Bonus for multiple scenarios!
To get you in the mood, here is my answer:
Quoting the man himself, Andrew Beyer:
From Picking Winners: A Horseplayer’s Guide, the beginning of the chapter on speed:
Among serious handicappers there are two major schools of philosophy. In one intellectual camp are the empiricists, who view every race as a unique problem to be solved by intuition and analysis. They evaluate horses’ records by weighing many factors and subtleties, and reject the notion that horse’s ability can be measured in any precise, concrete way. In the other camp are the rationalists, the speed handicappers, who believe that horse can be measured by how fast he runs. Speed handicappers perform various arcane calculations to translate a horse’s ability into a number. If an animal earns a figure of 99, he is superior to a rival that earns a 92. His age, sex, class, breeding and even his name is irrelevant.
By taking into account the different surfaces over which horses run and the different distances of thoroughbred race, they translated every horse’s performance into a single number that neatly defined his ability. The addressed the central question of handicapping: Who can run faster than whom?
I count myself as a empiricist according to his definition plus I also don’t believe you can boil a horse’s performance down to a rating that “neatly defines his ability.” Every handicapper has their own set of questions and their own hierarchy for those questions based on the circumstance (surface, distance, condition, etc). The question of “who can run faster than whom” is not even in my hierarchy let alone on top of it.
But I’m looking forward to all of your answers, because I know there have to be circumstances when even an anti-speed capper such as myself could leverage BSFs to my advantage. What say you?
The first handicapping book I read was Ainslie’s Complete Guide to Thoroughbred Racing, and the first track I handicapped was Suffolk Downs. I fall firmly into the empiricist camp, both by reading and experience. My primary interests are always class and form, but I find Beyer speed figures have uses as a secondary factor. In allowance races made up of primarily lightly raced 3-year-olds running through their conditions, or in allowance or stakes races where 3-year-olds are racing against older (particularly later in the year), I’ll look at figures to get a sense of how well each horse fits. There’s one pattern I especially like to see — a series of progressively rising figures, each number corresponding to a rise in class, competition, or distance.
I’m a rationalist empiricist.
I do think that the performance of a horse can be quantified, but I also think that it’s an inexact science not only because of all the variables involved within the race itself (weather, trip, weight, different run-up distances at different tracks, etc.) but also because of the potential for error in human judgment.
Taken at an extreme, I don’t think anyone would ever argue that Secretariat’s Belmont Stakes win is not a “faster” performance than what a bottom-level maiden claimer accomplished when fourth in his career debut going six furlongs at Evangeline. The “how much faster” is where conjecture would take place, but we don’t need to deal in such absolutes when handicapping a horse race.
Yes, it can be a game of noses (our version of inches), but to paraphrase Dana’s Beyer quote, in the end I just want to know who is fastest. If the odds compensate me for the margin of error then I’ll bet.
One of the biggest mistakes people make when handicapping races is asking the question, “Who will win” rather than “Who can win,” and it is the rationalist that Beyer speaks of who will look at the numbers and be able to determine who is too slow to contend.
And that’s where the speed numbers figure most prominently in my handicapping. Adjusting for perceived biases across region, surface, and distance, I can get a pretty good feel of who belongs at this level or against this competition.
I’d put myself in the empiricist camp but as a loyal reader of the DRF, it’s impossible to look past the ubiquitousness of Beyers in the form. With the bold numbers smack in the middle of the running line, I think most form readers find it impossible to look past them too. I also think, at the tracks that I play at least, that ubiquitousness is reflected on the tote. The Beyers pretty much own the tote at ovals with a traditional main track.
So I look for holes in the speed figures. Every figure can’t be bad, most all of them are actually fine as the relative value that they are. But if I think something looks a little off or doesn’t fit, I’ll ask myself a couple questions.
How was the figure earned?
Was the figure in question earned with a dream trip against lesser than today’s competition? Who was the competition? These are questions of class. Much more nuanced and they do require more digging and a better memory of past races and horses to answer. Digging is usually rewarded in handicapping.
Where was the figure earned?
There’s so many tracks, so many shippers and so many races it becomes obvious to spot the patterns of regional biases in speed figures. They’re supposed to be comparable between tracks but I just don’t find that true at all.
What surface the figure was earned on?
Turf and synthetics speed figures are unreliable. Best ignore them than to even take them with a grain of salt.
Using Beyer’s numbers made my handicapping life easier. That’s all there is to it. I used to do all the work myself, barely audible under an avalanche of chart books. I wish I hadn’t thrown out my trainer trainer- and trip notes – artifacts for the degenerate museum (mausoleum). Numbers were my weak point but I hated using other persons’ work; Andy, however, won my heart with his prose and hubris and brotherly three-fisted playing – HIS numbers I would use. Half of the runners (at least) are eliminated, shippers less inscrutable. Money time, if the numbers are fairly close, I’ll use current form over trips (replays), bias, hot trainer/jock, etc. It’s an honor to be involved even distantly with Andrew Beyer.
I use BSFs as directional only and I doubt there is anyone out there, including Beyer himself, who simply makes choices solely on figures.
For me, the use of these figures helps me gauge decisions by trainers, helping to answer the question of why a horse is in a race. If there are seven horses in a field and three of them have run relatively similar races recently as to the one coming up, it will seem obvious to me they are suited for each other from a competitive/class perspective. However, if two of the remaining four have drastically lower figures, I will spend some extra time on those horses, examining why they were pointed to the race (as Chris said, digging is usually rewarded in handicapping). It is here where I often discover value. As with Jessica, nothing makes me happier than a horse with recent steadily increasing figures corresponding with jumps in class and/or competition. Then, I immediately look for solid work patterns that are obvious preparations for the upcoming race.
Interesting, and a few things stand out to me but let’s take them one at a time. At the very least you each hinted at situations where Beyers might not be helpful. I know this subject was was discussed on Twitter and probably elsewhere, but I think it’s a good example to explore a couple of topics.
What do all of you make of Beyer discrepancy between Blame and Quality Road prior to the Whitney? Blame‘s previous three figs were all modest: 102 in the Clark, 101 in the Schaefer and 103 in the Stephen Foster while Quality Road‘s were 103 in the Hal Hope, 121 in the Donn and 112 in the Met Mile. Clearly, even if you’re factoring Beyers just a little bit, you probably landed on Quality Road, yes? Not only did Blame win but he got a 111 for a race that run relatively slowly.
It didn’t seem to me that Blame radically improved as much as it did that Quality Road didn’t put in a full effort, but the if I were looking at that fig and didn’t know much about either horse I would think Blame all of a sudden turned into a great horse. Or, is this a good example of taking BSFs in combinations with other factors, such as form cycle? Admittedly I have no interest or patience in the math involved in these calculations, but I’m wondering what all of you make this as you’re clearly more attuned to BSFs than I am.
As for Blame and Quality Road, again these numbers are directional only. I often throw out what I consider an aberration. In this case, the 121 would have been tossed. I wouldn’t have thrown out the race while handicapping, just the figure. As a result, and after looking at the PP, I fully expected Blame to run a solid race. Solid enough to beat Quality Road? Sure, but I’m not foolish enough to throw out Quality Road in my wagering strategy.
I guess I’ll answer but please be kind when (if!) disseminating this in terms of trying not to make things sound too poor out of context.
That said, a few things I read on Twitter from (of all people) @MarkAHorn and another gentleman emphasized that pars are just as important when comparing figures as are the numbers themselves, and I sort of buy into this.
Crist himselfâ€”when defending a particular figure on his blogâ€”is fond of noting that the number for so and so couldn’t have possibly been (any lower) or (any higher) because than all the horses behind the winner would have run (the slowest races of their career) or (the highest races of their career). That provides a pretty easy jumping off point for going into a race pretty much knowing the range of figure you’re dealing with in terms of what the winner will receive.
So, in the case of Blame, the low 100s are actually pretty good. I’m not sure that his BSFs for the Pimlico race or the Foster were the ceiling for either of those races, but I do think he earned an exemplary figure in each attempt. That idea alone puts him in Quality Road‘s ballpark since I think Blame had faced tougher competition this year.
Again, the idea isn’t always who will win but who can win, and given the odds on Blame it was pretty easy to put him in the overlay category if you liked him (I say this thinking Quality Road couldn’t lose but that had more to do with me buying into the Pletcher (and #EastCoastScribes?!!) hype than the figures.
I think this angle is especially powerful with maiden races. Horses who win at Turfway or Ellis just don’t get the same figures horses who win at Keeneland and Churchill do, but when a MSW winner comes to Keeneland or Churchill from Turfway or Ellis sporting a BSF in the 70s then I know this is a serious racehorse and likely better than a horse who earned a 70 at a major circuit.
I was halfway to replying to this and then I re-read it. Too snarky and redboard-ish. I torched the whole thing.
The general gist is this. That 121 Donn number is ridiculous. Sure, it’s probably legit but that race was *long* over at the 3/8ths pole. I watched the replay again the other night and it came off like a public workout. Then I looked at the chart of the race and who he beat and it’s a pretty bad lot for a Grade 1. I remember handicapping the race and trying to find a horse who could beat him and even I, staunchly anti-chalk, keyed Quality Road on top of trifectas. Like Ed says, there was no other horse in that race who *could* win when I handicapped it.
With that in mind, Blame was obviously serious competition for Quality Road in the Whitney.
Couple other points building on what Ed said.
– There’s a huge gapping hole in regard to defending the Quality Road‘s Beyers in the way Crist normally does. Musket Man has earned his 3 best Beyers in defeat to Rachel Alexandra and Quality Road. It’s just tough for me to take that seriously.
– I’ve never understood *why* the Beyer peeps go so far out of the way to combat the fig haters. I don’t get it. Never have. They make *huge* adjustments to their figures all the time across the board. So what’s the big deal when someone points out that there could be a flaw?
After re-thinking what I wrote and reading Chris’s contribution, I REALLY want to hone in on what I said about the maiden races because it’s been a hugely profitable angle for me to take seemingly mediocre figures on either synthetic or lesser circuits and promote that horse’s chances when stepping up to a major circuit but not stepping up in class (i.e. MSW to N1X or MCl to N2L).
The other issue I thought of that relates to the maiden issue as well as to something Chris pointed out relating to Musket Man is that I trust figures much more for winners than I do for also-rans.
In The Odds Must Be Crazy: Beating the Races with the Man Who Revolutionized Handicapping the Lens (Ragozin and Friedman) talk about how Ragozin Sheets times multiple horses in a race to get more accurate figures for the whole field, but I’m not sure the BSFs use that method, meaning that every figure in the race is relatable to the winner’s and because of the variables I talked about in my initial post (weight, wind, racing wide, etc.) there is thus more margin for error.
A 75 BSF by a winner is far more impressive to me than an also ran who earns an 80, and typically the higher the BSF for an also-ran the worse I expect them to run, which typically works out because those horses get over bet.
Thoughts on accuracy of BSFs for winners versus also-rans?
The question about Quality Road and Blame and Chris and Ed’s replies gets to a fundamental weakness in Beyer speed figures — the numbers tell you nothing about pace or conditions. A horse that “runs away” with a race, especially against weaker competition, will often get an inflated figure. It’s most noticeable in graded stakes, but it happens at all levels.
Not to redboard, but Quality Road has always been what I think of as a typical one-dimensional dirt horse — he doesn’t have the ability to accelerate in the stretch. He looks like he does, because he’s capable of more speed throughout a race than most of his competition, but in the end, he slows like most — just at a lesser rate. That makes him vulnerable to horses that can accelerate, as Blame did in the Whitney. Beyers reveal nothing of this scenario.
Zenyatta specifically (and synthetic and turf races generally) illustrate Jessica’s point.
Zenyatta has run everything from downright SLOW numbers for a Grade 1 race (94 in the Hirsch I think) to world-class numbers (110 in the 2009 Breeders’ Cup Classic). We all know her styleâ€”keep up early then blow by them in the end. She’s running her race the last three furlongs (similar to top turf horses), and BSFs (and to be fair other figures) are always going to have trouble with those types of scenarios.
Great stuff everyone, and feel free to keep this thread going if anyone has something else to add, but I want to put another question out there as well…
We’ve addressed where BSFs can help identify an up and comer, how they help sort out who doesn’t belong and where one might want to take them with a grain of salt. The last two answers have started to focus on something else I’m interested in: Identifying false favorites. I’ve had discussions with a few of you on this matter I know there are a few more scenarios that one might want to keep an eye out for.
Two scenarios come to mind —
Older horses with established form who are starting off a new career-top speed figure. Rarely does that number represent a sustainable breakthrough, but the horse will be bet like it does.
Ed’s observation re: the accuracy of BSFs for also-rans is right on, particularly in maiden and juvenile races.
This sort of goes along the lines of trusting a winner’s figure more than an also-ran’s, and it probably has more to do with handicapping a horse’s form more than his actual speed, but I definitely am weary of horses whose career debut is a big number.
It’s not that I don’t trust the number, it’s that I don’t trust the horse to repeat it. They are almost always not only the favorite but also ridiculously over bet. If I have any other reason to like another horse in the race, then I’ll definitely include it in pick Ns and try to beat the odds on fave.
90+ on the BSF scale seems to be a good threshold for career debuts. Note that Boys of Tosconova got a 102 in his second start and “bounced” to a 90 next out. I can’t wait to bet against that horse with both fists at Churchill. 7-9 weeks between each start? Ugh.
Excellent points. Ok, it seems like we’re winding down here, is there anything else that anyone wants to point out? Freestyle time!
There’s been a years long fascinating discussion over at Formblog on the west coast versus east coast Beyer figures. I’ve always thought both sides of this discussion to be slightly over charged for my tastes but yesterday Illman posted this data which is very striking.
The quality in California is nowhere near what it used to be but they’ve run two Breeders’ Cups at Santa Anita during the synthetic era, you’d think there would be some top performances in there. You don’t need to have a degree in statistics to know there is something awfully wrong with this distribution even if some may say, perhaps rightfully, that figures don’t work that way.
From a wagering perspective, this cat has been out of the bag for quite awhile. All one needs to do is play California races for a little while to see tried and true speed figure handicapping methods are irrelevant.
333 Beyers of 100 or greater; 38 at 110 or greater; 2 at 120 or greater
320 Beyers of 100 or greater; 9 at 110 or greater; none at 120 or greater
420 Beyers at 100 or greater; 78 at 110 or greater; 3 at 120 or greater
410 Beyers at 100 or greater; 28 at 110 or greater; none at 120 or greater
214 Beyers at 100 or greater; 38 at 110 or greater; 3 at 120 or greater
167 Beyers at 100 or greater; 17 at 110 or greater; 2 at 120 or greater
Santa Anita (2003-2006):
429 Beyers at 100 or greater; 42 at 110 or greater; 1 at 120 or greater
Santa Anita (2007-2010):
162 Beyers at 100 or greater; 4 at 110 or greater; 0 at 120 or greater
Del Mar (2003-2006):
143 Beyers at 100 or greater; 18 at 110 or greater; 1 at 120 or greater
Del Mar (2007-2010):
135 Beyers at 100 or greater; 2 at 110 or greater; 0 at 120 or greater
277 Beyers at 100 or greater; 37 at 110 or greater; 1 at 120 or greater
110 Beyers at 100 or greater; 1 at 110 or greater; 0 at 120 or greater
And there you have it have it folks, some interesting insights and ideas on Beyer speed figures and how to factor them. A big thank you to our contributors for their participation!