Monday, 28 February 2011

Equine Flow Psychology I

When I first started analysing sectionals from the US 15 years ago, I found the task tremendously interesting. It felt like I was learning things about racing that few people in Britain knew or understood.

I used to listen to experts on television here describe front-running winners as having "done it the hard way". Nowadays, the same performances would be termed "benefiting from an easy lead".

But when races had featured duels early on, I would come across horses who had run six furlongs like this:

:22     :45     1:12

and next time, in far less contentious situations, they would cover the same course and distance like this:

:23     :46     1:10

It happened time and again. There were plenty of examples. Nothing had changed about the horse between its two races apart from the way it ran the race.

If you read any part of my series Pace: the final frontier, you will know exactly what is happening here. (And, though I did not recognise it at the time, the whole concept was part of the orthodoxy of US racing analysis.)

By covering the opening quarter-mile a second slower, a horse could run two seconds faster for the full race distance. This was an astonishing difference. At once, it opened my eyes to the power of pace.

Had a horse like this been assessed by traditional methods of doing ratings in Britain and Ireland, handicappers would have it improving by the best part of 30lb!

As you can imagine, the prospect of being able to actually project similar amounts of improvement made me feel as if I could turn base metal into punting gold.

So, I got to work and calculated "too fast" opening quarters for every course and distance which was televised in the UK. I intended to back horses who had conformed to this idea and watch the profits come in.

Of course, it didn't happen. It was only Fool's Gold that I had discovered. I should have realised that someone else would have got there first.

I did find some nice winners, mind. One night at the Fair Grounds, I nailed a 37-1 winner that seemed so obvious it was untrue. But there were plenty of horses who ran disappointingly, and there was something worse still.

When I looked at their performances as a group, about 50% had indeed run faster times than before. But by only three or four lengths - not enough to turn dismal losses into anything approaching wins.

About half the number or 25% of the sample - including my 37-1 winner - had jumped up considerably, just as I expected. Okay, usually the payoffs were not great because many were dropping in class.

But 25% had not even run a faster time. Not by a fifth of a second. These horses simply had to run faster, according to my fundamental rule of pace.

I was shocked. And when I looked at the individual cases, I was more confused still.

A lot of these horses were front runners who ran hard early to set up a lead, then tired badly in the straight and were caught in the final furlong. Next time, their jockeys did indeed try to ration their speed and run the opening quarter-mile more steadily, but they seemed to tire in exactly the same fashion and did not improve their times.

Like any bad scientist, I decided to disregard these horses from my sample. Instead, I proceeded to work out an equation which linked the final time of a horse to the way it had run each quarter-mile.

It pleased me greatly that this seemed to work pretty well. But still there were horses who did not obey the "rules".

I rationalised that many of these horses were probably unsound or had been treated with illicit medications. I was probably right, but it did not remove the nagging doubt from my mind.

Something was happening that I did not understand. My new-found "genius" was flawed.

It was then that I read James Gleick's superb book Chaos: Making A New Science (Viking Penguin). After only 50 pages, I went back to my data and began to look at it in a whole new way...

Friday, 25 February 2011

Bounce Theory

In Britain and Ireland, a horse having its second run after a lengthy absence is sometimes referred to as a "bounce candidate". The concept is that freshness may have caused overexertion last time, leaving it sore and vulnerable to running poorly today.

Though some horses undoubtedly fit this pattern, the evidence may be just circumstantial. Horses with a history of absences from the track often have chronic physical problems which will resurface sooner or later. But they may not be affected on the second run back, it could just as easily happen on the third run, fourth run or later.

Not surprisingly, those who have investigated this particular "bounce" have found statistical support hard to find.

But the bounce theory as a holistic concept has little in common with this particular usage. Instead, it may be thought of as just a theoretical basis to understand the variance of racehorse performance over time.

All performance figures are affected by the randomness inherent in a race. A horse may get the ground or pace which suits it one day, but the next it may end up drawn badly or the victim of bad racing luck.

Some of these factors are obvious just by watching the race. But there is also an undercurrent of randomness in the physical state of any living being - just by dint of the fact that its health and wellbeing vary from day to day.

While handicappers have traditionally made great effort to improve and maintain the accuracy of their ratings, not until recently have some begun to investigate how past figures actually predict performance.

Phrased more technically, these far-sighted analysts have asked the question of the optimal weighting to apply to each particular rating in the array of a horse's past performances.

How much does recency matter? To what extent should we favour the consistency of an inferior horse over the peak ability of a superior one?

Many of these questions can be answered by sophisticated statistical techniques and neural networks. These powerful methods lie behind the great success of some of the most technically minded punters out there.

But what can we derive from the data that is useful to apply for those who don't have access to such information? How can we conceptualise the potential for variance in a horse's ratings over time - without using correlation coefficients or regression equations?

The bounce theory.

Because of the randomness of races and the varying physical state of racehorses, the performance rating of a horse in any particular race is just one expression of a whole array of figures it might have produced on the day.

So, the chance of it reproducing the figure depends on the the degree to which it was representative of the horse's mean output - at that specific point of its career.

When we look back at a video of the horse's last race, or consider the facts surrounding it, it may be totally obvious that its performance was a fluke. The horse may have benefited from a superior ride or raced on a favourable part of the track or lucked out in some other way.

But we don't necessarily have to know this. Neither does it have to be obvious. Instead, we could look back at its previous performances and infer that the latest might be aberrant because it is significantly different in merit.

In these circumstances, it is reasonably to expect that the horse will regress to a mean-level performance more often than not - that it will return to running to a figure closer to that which it had previously established as representative.

It is hugely important to understand that the use of the word "regression" in a mathematical context does not necessarily  have a negative connotation. The bounce theory also applies to a horse whose recent run was poorer than reasonable expectation; in this case, we might expect it to "bounce back to form". (Or, strictly speaking, towards form.)

Expressed formally, the bounce theory may be stated thus: "The tendency of a racehorse's performance level to regress to its notional mean."

The key word in this description of the bounce theory is "notional". A racehorse's career develops differently according to many factors, of which its age and experience are the most influential.

So, its mean, or expected performance level cannot be calculated by taking the average of its last three runs, or some other clumsy stricture.

Instead, the performance level of racehorses in general will change over time according to an age/experience curve: younger, inexperienced horses get better, older ones are more consistent or begin to decline.

While neural networks and techniques such as Markov chain analysis enable us to reach an estimate of a horse's likely performance by mathematical methods, this can still be limiting. There are many influences which defy strict statistical interpretation, such as how the trainer regards it or whether today's race is a particular target, for example.

Instead of getting bogged down in data analysis, many less technically oriented punters can use the bounce theory merely as a guiding principle; they can simply think of a racehorse's ratings as a notional progression.

When any of its performances depart sharply from reasonable expectation, a "bounce" or regression can be expected. In other words, the evidence of a horse's most recent run can be put into its proper context and seen just as a single point on the "best-fit line" of its career trajectory.

Of course, many intelligent punters already use the bounce theory. They just don't formalise it as such, or refer to it by name.

In essence, the bounce theory is a framework which encourages you to think in terms of expectation. And not to rely too much on the evidence of a horse's recent race, especially when you believe it was not truly representative.

Pace: the final frontier - final Q&A

A massive thanks for all your questions, both in the comments section and on Twitter. I include the best of them here but there were several others that should have made it but I couldn't put in for Gone With The Wind-style concerns.
In particular, Colin was really unlucky not to be included. His points were either raised by others or answered by them.

* * *

administrator said...
Q: The problem with Wilkins book is that like many academics he becomes engrossed in the method rather than the end results. There is a comprehensive set of articles at for those interested in further exploration of pace and profitable ways of employing it. 

A: This was just one of a number of useful observations you made, Admin. I hope you don't mind me not including the rest. I have taken all your points on board.
Yes, as I said in my own review, I believe Wilkins' book Bioenergetics and Racehorse Ratings is stronger on the former than the latter. He implies as much himself.
I am glad you mentioned the excellent work that and its forerunner have done. I have no link with them whatsoever and am not here to do endorsements, but I completely agree that research both on their website and in their excellent book Improve Your Betting is well worth accessing for punters wanting to make practical improvements.
(I might also mention the website as another portal from which technically minded punters who don't have my obsession with the abstract can find much to nourish them.) 
For what it is worth, Admin, I am very much from the Wilkins' school. I am motivated to spend my life on ground-level theoretical research of pace and racing analysis because I have no confidence in the superficial methods generally employed. Moreover, I distrust received wisdom.
To be honest, I could not compete intellectually with the expertise of the guys in their particular field. Top-class stuff.

cavelloman said...
Q: How do they measure times from a rolling start in the US?

A: Gavin, the timing apparatus is triggered 30-35 feet in front of the stalls on most tracks, so the actual distance of races is slightly more than that advertised.
There are important differences from track to track and, crucially, even from one start to another on the same track.
Over 6f at Churchill Downs, for instance, first-quarter splits are very fast presumably because there is a more generous run-up distance than elsewhere. It is not important to know the exact distance of the run-up because it is inferred as part of the data when calculating standard sectionals.

Cormack15 said...
Q: I agree Bioenergetics and Racehorse Ratings is a fascinating book. I reviewed it on at the time of publication and, from memory, my concern was that the application of the methods allowed/relied on too much emphasis on subjective judgements. But I also recall Bob explaining in the book that the practical application wasn't his primary motivation nor interest. Perhaps the door is open, James, for a Wilkins/Willoughby collaboration.
A: I think we are all singing from the same hymn sheet, David.  I fear a Wilkins/Willoughby collaboration would draw rather too much on the former than the latter, but it goes without saying I would do it at the drop of a hat. You must remember that I am not fit to lace Bob's boots academically.

R Hills is God said...
Q: Ah, Bob Wilkins. A name oft mentioned in hushed and clipped tones in betting shops up and down the land [Britain]. Every time a gamble is landed, someone starts a rant about insiders before a more mature punter pulls them to one side and tells them: "You do realise the stable probably didn't have a penny on it. People just assume that they have backed it when the majority of market moves are caused by Wilkins' money."
I've met every shadowy figure in the betting world that has controlled the ring with a rod of iron, or have met someone who has met them, save for one: Bob Wilkins. Everyone claims never to have met him. He's the Naylor's Naylor.
It would be an incredible feather in your cap James if you could get the great man out of hiding and have him hold forth over a few cold ones.

A: Glenn, this is another special contribution. You are so right. And my particular cap has plenty of room for feathers. The "Naylor" to whom you refer is, I believe, the shadowy Pat, the only man truly deserving of the word "genius" as applied to betting. 
I believe that Wilkins has, in effect, let the cat out of the bag. It is now our job to make sure the bag ends up in our office.

Cormack15 said...
Q: You mentioned "regression to the mean" in relation to bounce theory. I'd like to hear more on your thoughts on that.

A: David, if you continue to read this blog, you have no choice! Regression is the most useful and oft-used technique in modern-day statistical analysis of sport. It is at the core of econometrics, the mathematical disciplines which have changed the way we understand the world, not to mention sport.
I will be presenting my theoretical approach to bounce theory (not that second-run-of-a-layoff nonsense) in a blog coming up soon. And to illustrate a working example, I will be drawing on data from my second love, Major League Baseball.

Mark Milligan said...
Q: The books that James mentioned are both superb but I also strongly recommend Figure Handicapping by James Quinn and Modern Pace Handicapping by Tom Brohamer. Both works are exceptional in my opinion.

A: Mark, I agree that they are both useful and informative. But I am not so keen on readers simply adopting their pseudo-scientific approach to pace and analysis. Using Wilkins and Carroll gives you formal understanding of the core principles, on top of which you can layer your own approach tailored to your own particular intelligence. 

mick.ingleson said...
Q: As you have used the Classic as an example, can you compare Blame's sectional timings against Zenyatta's? Was she just a fast finisher, or were the others tiring?

A: Hopefully you can trust me to give an answer without proof on this one, Mick.
The way the Classic was interpreted on the television left me a little frustrated. Admirable and talented though she is, Zenyatta is no wonder horse. She was very good indeed on the track - but not great in terms of absoloute merit. 
Her record and her accomplishments do perhaps deserve the term; either way she should be accorded utmost respect. But I get a shade frustrated when analysts cannot separate a horse with a brilliant record from a brilliant horse.
Zenyatta lagged behind in the Classic entirely because of her own attitude. Mike Smith always rode her that way because it was the best way to maximise her energy while staying within the bounds of her psychological needs.
When a horse is so narrowly beaten, it is impossible to say the jockey could not have done anything to win the race, but such a unilateral approach to analysis is always deeply flawed.
Zenyatta's amazing record said much about her durability and, indeed, her talent. But her style of running and the risible competition among Californian females led to the appraisal of her merit becoming somewhat overblown.

Robert said...
Q: Good stuff, you did not explain why Secretariat managed his uneven performance in the Belmont over 1m 4f though.

A: Robert, I am sorry that was an oversight I will put right. The clue to it is something I am calling "Equine Flow Psychology" which I believe explains the presence of some outlying sectional data.
If you wish to get a jump on this for yourself before I blog it, research its human equivalent and apply your intellect to the particular dynamic of equine athletics. 
Once again, apologies for the omission. Sadly, it is going to happen again because I am excitedly juggling too many plates in the early going here at TFNL.

jackiejameson said...
Q: There is room for discrepancy in analysing sectional times even within the same race meeting, as the Breeders Cup meeting at Churchill Downs illustrates. 

Morning Line's sectionals in the Dirt Mile were exceptionally fast up front and set the race up for the closer Dakota Phone, yet he was nailed only on the line, whereas the four horses up front in the Classic stopped as if shot having gone through 6f around 1:11 / 1:12 which was not especially fast with reference to the Dirt Mile sectionals. 
And the reason for this? The dirt track was watered after the Dirt Mile and before the Classic on foot of representations from horsemen about its condition and the effects of that are readily apparent from the sectional times for both races.
A: A very interesting submission, JJ. Your point about the watering is a valid observation but I don't agree with your inference of its overwhelming effect.
I believe the mistake you may be making is assuming that the sectionals of a race over one trip can be compared with those of another.
According to both my database of Churchill Downs and the useful publication HorseStreet Pars, a Graded Stakes-quality horse capable of running the Mile in 1:35 would achieve a split time of 1:38 for the first mile of the Classic. And the parallel split times (what I call standard sectionals) for both races show a similar disparity.
You only have to look at the way an opening half of :47 absolutely bottoms horses in the Derby, yet can be happily sustained by cheap claiming horses over the one-mile trip at Churchill.
In real terms, therefore, the horses up front in the Classic were running a lot harder than you imply. And that is probably the main reason they tired, in addition to the concept of "pace pressure" caused by a multi-horse scenario up front.

cavelloman said...
Q: It looks to me that the bends are having an effect on the times. In the Dirt Mile and Classic the sectional times on the back straight 6-4 furlongs were faster then the previous splits which was achieved on the first bend.

A: With regard to the last question...Bingo! See what this blog might achieve if I can continue to attract so many great contributions? The bend may not be the only reason for the disparity between the Mile and Classic splits - I accept that watering may have had some effect - but this is a good bit of purposeful thinking, Gavin. 

Denis Beary said...
Q: Have you attempted the task of producing standard sectionals for different class levels for any tracks in the UK or Ireland? If not is this because you think doing this accurately from video recordings is not possible? If yes have you found the figures to be the Holy Grail of punting?

A: Denis, you have shrewdly answered your own point. We don't have the accurate data necessary. While we can use paths, markings on the rails and, indeed, the furlong-poles themselves to derive workable sectionals and learn an awful lot, doing so as a holistic exercise for British racing is, sadly, impossible. 

Charlie Dickinson said...
Q: "The Match Up supercedes everything."

A: Not in my view, Charlie. If you are referring to the interplay between the running style of horses and the race within a race, it is extremely important.
I like a lot of your stuff and I do realise your phrase is designed to be provocative through its succinctness, but I don't like such rule-based analysis. 
I believe we should be esoteric in our approach. In pace and in life.

Tuesday, 22 February 2011

Pace: the final frontier VI

As I present this final part of the series, I want to thank everyone who has made the effort to follow along. Nowadays, there is so much going on in our lives it can be tough to find time to embrace new ideas.

My ambition for Pace: the final frontier was to get you thinking. I deliberately wrote it in conceptual form for this purpose; the best way to engage people in a subject is to present its basic ideas in a way which individual minds can develop for themselves. I hope that has been achieved.

As the weeks and months of the Flat season progress, I will be relying time and again on these first principles to develop ideas specific to races and horses in the current spotlight.

Inevitably, a few of the examples I draw will be from the US, France, Australia, South Africa and Hong Kong - in other words the countries whose punters are enlightened by sectional times.

As I write this, it is still hard to believe that the UK and Ireland are missing from that list. Those who say that sectionals cannot add anything to British and Irish racing cannot know what they are missing. 

When I first started watching racing 25 years ago, it was extremely rare to hear even a horse's final time mentioned on television. But, tune into Racing UK now - and sometimes even At The Races - and you will hear cogent, intelligent analysis of times on a daily basis. 

But, we should have moved on from that 20 years ago. Sectionals enable us to understand the limitations of final time as an objective assessment of performance. To repeat that mantra: it is not how fast they run, it is how they run fast.

Given that British and Irish racing doesn't have sectionals and is unlikely to get them in the near future, I will show you it is possible to apply the concepts of pace developed here without them. All that when the Flat season on turf finally arrives.

* * *


5.2 Final definitions

As we conclude, here are three important definitions which have been established in the text. If you take nothing else from the series, merely a conceptual awareness of the following should nourish your understanding as you develop ideas about pace suitable for your own application.

Even pace is merely the subdivision of a race into sections of distance and time which represent the minimum variance in average speed. 

In the US, races are timed from a running start, so even pace for a 6f race may be represented by three two-furlong sections of 23sec, for instance. In countries where times are recorded from a standing start - such as the UK and Ireland - even pace for a 6f race may instead be represented by three two-furlong sections of 24.7sec-23sec-23sec.
This, of course, is not constant speed but the closest thing to it, given the implications of a standing start.

Ideal pace is the real-world application of even pace, taking into account not only the effect of features such as bends and hills, but the different energy types available to a horse over various distances. 

Sprinters have to run slightly faster than even pace early when powerful, anaerobic energy is available to their many fast-twitch muscle fibres. But this effect rapidly fades with distance.

"Ideal pace is a proxy for optimal energy use," Glenn Alcoe puts it.

Ideal pace for a 6f race at Belmont Park, for instance, might be represented by two-furlong sectionals of 22.8sec-23.5sec-23.2sec.

The first section is the fastest because the proportion of total energy derived anaerobically is at its highest; the second section is the slowest owing to horses having to negotiate the bend; the third section is slower than the first but faster than the second because it is straight but the horse is now much more dependent on aerobic energy.

Standard sectionals are the fractions most commonly run by horses achieving fast times over a particular course and distance. They vary from ideal pace because of the tactical demands of a race, particularly the need to run hard early in US racing in order to make a bend.

Standard sectionals enable us to profile the energy requirements of a particular course and distance, and sometimes specifically a race like the Derby at Epsom or the Kentucky Derby.

In galloping tracks in Europe, standard sectionals are usually very close to those represented by ideal pace. But at tracks in the UK with tight bends or short straights, standard sectionals are again tilted more towards early pace.

5.3 Recommended reading  

There are few good books on the technical analysis of racing and energy use. They are not specifically about sectionals but give you the ability to reach that understanding. The two that stand out to my mind are:

Handicapping Speed by Charles Carroll (The Lyons Press). A wonderful, esoteric US book with sections that are so good they nearly blew my mind. I used to go about for the rest of the week with his words circling in my head. Carroll led me to become interested in Chaos Theory.
When I first read this book, it made me think how little intelligent analysis had penetrated horse racing. It convinced me that spending my life thinking about racehorses was worthwhile because of the beauty of intelligent expression.
Whenever I feel a bit down, I reach for this book. Parts of it are so well expressed it is bewildering, almost emotional to read.

But the No 1 all-time in my view...

Bioenergetics and Racehorse Ratings by Bob Wilkins (Overdee Press; I have never met the author but I would like to. When I read this book, I became terribly depressed: I saw it as the book inside me, the only one I ever wanted to write but didn't have the balls or the brains.
Wilkins' magnificent mathematical adventure through human and equine athletic performance is not for those frightened off by equations. 
But, as a friend told me on reading it: "You don't need to understand maths, you can just trust Bob."
I feel the book veers a little into subjective territory with its treatment of ratings near the end, but that is like saying that Jayne Torvill had a zit on her nose during Bolero.

5.4 Conclusion

Hopefully, this series leaves as many questions open as it has answered. That would be great.

I will be attempting to answer many of these - including that lingering one about Secretariat and the Belmont - in separate posts which are self-contained. Though not much of a populist, even I realise that readers are often dissuaded from tackling posts described as "Part XXVII", for instance.

Next time, I will answer any remaining questions the series has thrown up.

Monday, 21 February 2011

Pace: the final frontier V

In this penultimate episode of Pace Wars, I establish how standard sectionals vary according to both track conditions and the distance of the race.

The final part will underscore the important definitions made during the series. 

Plus, I will belatedly reveal the name of the book I consider half-an-hour the best for those interested in the technical analysis of the sport.

* * *


4.9 The effect of track conditions

The predominant racing surface in European racing is grass, whereas dirt is traditionally favoured in the US. Now, synthetics are making an impact on both sides of the Atlantic.
Not only does the composition of these different surfaces exert an influence over the way races are run, but their condition is also highly influential.

Grass tracks can vary tremendously in moisture content. This makes grass a multi-dimensional surface as far as pace is concerned - the moisture content plays a big part in determining what type of horse is successful and which style of running is favoured.

In other words, different going conditions - from firm to soft - cause a shift in standard sectionals - the most common way that fast times are achieved over a particular course and distance.

As turf becomes soft, acceleration starts to count for less. A horse's cruising speed tends to take a lot more out of it because its stride becomes far less energy-efficient. Standard sectionals therefore describe a more marked pattern of deceleration.

But, as the surface gets really heavy, the pattern of races becomes chaotic and optimal energy use becomes blurred. Horses who lag behind through having no early speed can make big runs at the exhausted leaders.

Quantitative pace analysis on turf tracks in Europe graded soft or heavy is highly dubious, in my experience.

Dirt tracks in the US are more consistent in speed and moisture content. But there are still variations which impact the run of the race.

The most dramatic are the sloppy or muddy surfaces which follow heavy rain. Again, the finer points of energy use are now subjugated by the base need to get out of the kickback. So, early pace is heavily favoured.

But when such a track dries out to good, the slant on pace can soon change.

Synthetics are closer to turf than dirt in their influence over the shape of races. They tend to put the emphasis on a horse's finishing speed.

Indeed, the higher final-quarter velocities of some synthetic surfaces magnify the effect of trouble-in-running. A horse whose momentum is interrupted faces a costly loss of kinetic energy and has little time to recover its velocity.

Despite the broad similarities between the pace of turf races and synthetics, stallions who have proved prepotent on the grass may not prove so effective at getting high-class stock on synthetics. 

Stallions like Sadler's Wells, Galileo and Montjeu impart valuable stamina to their stock which is valuable on the undulating tracks of Europe, particularly when the going is soft, but this blue-chip endurance does not seem to play such a part on synthetic tracks.

Additionally, although calculating standard sectionals on the various synthetic surfaces is no different to the task on turf, the way races are run shows much greater variance.

Perhaps jockeys still don't have a feel for synthetics. Either way, these tracks carry particular difficulties for the analyst.

5.0 A feel for how standard sectionals vary by race distance

Consider the following sectional times from three dirt races held at the 2010 Breeders' Cup at Churchill Downs. The splits given are those individual to the winners, not for the leader.

Classic (1m 2f)  Blame          24.50     24.53     23.74     24.58     24.94
Dirt Mile (1m)   Dakota Phone             24.75     22.40     23.60     24.55
Sprint (6f)      Big Drama                                   21.34     23.21     24.50

Now look at that same data with the fastest (min) and slowest (max) quarter run by each horse:
                                                              min       max       diff
Classic (1m 2f)  Blame                       23.74     24.94     1.20
Dirt Mile (1m)   Dakota Phone            22.40     24.75     2.35
Sprint (6f)      Big Drama                    21.34     24.50     3.16

I could have employed a more technically correct method, but hopefully this slanted example makes the point: the longer the distance of a race, the more that standard sectionals converge on even pace.

In other words, sprinters tend to run much more unevenly than stayers. We have already touched on part of the reason for this on dirt - the need to get to the first bend in front. But there is a more fundamental explanation which applies in a universal setting.

5.1 Different types of energy available to the running horse

This is a wide-ranging topic which I will deal with more expansively in a future blog. Like humans, horses employ both anaerobic and aerobic respiration to meet their different energy requirements through a race.

Anaerobic energy is explosive and empowers a thoroughbred's fast-twitch muscles as it overcomes inertia in the first part of the race.

But this energy source is dirty and soon becomes limiting. As the race goes on, the horse depends on aerobic energy in the attempt to maintain its momentum or even to accelerate.

The longer the race, the higher proportion of total energy is derived aerobically. And this relatively clean source is best used by a horse's slow-twitch fibres at a steady rate. So, it is manifest over longer distances by a horse running most efficiently at even pace.

In sprints, the greater dependence on anaerobic respiration leads to a a horse running its fastest sectional times in the early stages. Typically, a sprinter bursts from the gates, soon hits top pace and decelerates thereafter.

But this pattern is far more common in sprints run in the US than Europe. Sprinters on tough, demanding turf tracks in Britain and Ireland tend not to run flat out early, particularly when the course has no bends.

For those interested in comparing thoroughbred energy use to humans, note that aerobic respiration is a more significant component for horses over all distances commonly run. There is no thoroughbred equivalent of the Olympic 100m.

Saturday, 19 February 2011

Pace: the final frontier IV

Note: A few minor tweaks have been made to the original version of this post to clarify some of the terms I am using to describe sectional times in this series.

This followed an extremely helpful suggestion by Glenn Alcoe in the original comments section. As this is now fully integrated in the revised post, I have deleted his comments only to save others reading them becoming confused.

In the next article, I will conclude with a clear definition of terms such as even pace, ideal pace and standard sectionals. 

Thanks for reading. 

* * *


4.6 The influence of track layout on ideal pace

The question I left you at the end of part III was to explain the fractions for those Churchill Downs claimers over six furlongs. 

To refresh, the average individual sectionals for winners typical of the $10,000 level in 2001-3 were:

22.2      23.6      25.0

This pattern of deceleration, I remarked, was nothing like the way we have established that fast times are usually achieved. Most commonly, horses achieve optimal time by running at even pace, allowing for the physical features of a particular course and distance.

Over six furlongs at Churchill Downs, even pace for a winning time of 1:10.8 would be represented by quarter-mile fractions of:

23.6      23.6      23.6

But, after an average first two furlongs, a horse running its race in this fashion would find itself 1.4sec (23.6 - 22.2) behind the point where winners are usually running. According to the convention that a length is the equivalent of a fifth of a second, this time-lag would see it buried in the pack seven lengths behind.

In this position, the poor beast in question could be in between horses and being pelted with kickback. Doubtless, it would have to wait for room or have to be angled wide to make its run.

Either scenario is bad news: running wide on tight tracks involves serious loss of ground, but waiting for room, or being hampered, is even worse.

Any interruption to the momentum of a horse running evenly can be more costly than the extra effort required to stay out of trouble.

But, a jockey cannot afford to be too aggressive early. If too many of his rivals want the lead, the energy required of his mount to get to the front can easily become exhausting.

So, one definition of standard sectionals could be: the tempo which balances the extra energy spent by running uneven splits with the energy saved by having the run of the race.

To clarify: a horse running over six furlongs at Churchill Downs has to run fast enough to get a favourable tactical position, but not too fast so that it has nothing left for the drive.

And the way this is achieved most commonly is described by the standard sectionals for that course and distance.

These might be expressed generally for all horses - as standard times are in the UK - or specific to a class level - as par times usually are in the US.

Earlier in this exercise, for instance, I revealed that the standard sectionals for a $10,000 claiming race over six furlongs at Churchill Downs were:

22.2      23.6      25.0

which expressed as elapsed time is:

22.2      45.8      1:10.8

4.7 Profiling the track with standard sectionals

The process described above enables us to take a leap of understanding from that provided by final-time speed figures like those of Beyer and his associates in the US or Timeform in the UK. In addition to an objective measure of how fast a horse covered the distance of the race, we can now evaluate how it did so.

The importance of this consideration is captured by a popular aphorism among horseplayers in the US: it is not how fast they ran, it is how they ran fast.

Standard sectionals for a course and distance are calculated by considering a suitably large sample of horses who have achieved their final-time potential. 

In calculating the most common way they achieve fast times, we are said to profile the track.

In Europe, this exercise would be even more useful for horsemen. It may soon be apparent to a jockey at Churchill Downs that he needs to look sharp early, but what about the best tactical approach to running a mile and a half at Epsom or seven furlongs at Goodwood?

As this is a more complex task to learn from experience, a jockey or trainer with this information for every course and distance in Britain would have a huge tactical advantage.

4.8 Implications of different track surfaces

But the shape of the track is not the only consideration in establishing standard sectionals. Its surface arguably exerts just as much influence.

Returning to our Churchill Downs example, I am going to turn to the excellent answer given by my correspondent Robert in response to the question I left you last time:

"The reason for the shape of the sectional times at Churchill is mainly the surface. Early speed is king in dirt races because it's hard to quicken, and any lost momentum due to trouble is magnified," he wrote.

As Robert rightly says, the traditional dirt surface of the US tends to shift more under a horse's hoof than either grass tracks or synthetics like Polytrack or Tappeta.

As a result, we say that dirt surfaces generally "favour early speed" and "don't play towards acceleration". Whereas the opposite applies to turf and synthetics.

Expressed more technically, final-quarter velocities typically are not as high on dirt as other surfaces.

* * *

Being mindful to heed your views to make this journey into pace steadily, I am going to stop there for now. The next part will discuss the influence of different track conditions - fast or slow.

And we will get to the most interesting factor of all - how ideal pace varies according to the distance of the race. This factor has profound implications for the subject of thoroughbred pace as a whole. You might want to give it some thought, if you have time.

Thanks for reading. Please leave any questions you may have, or any clarifications I may add, in the comments section underneath.

Friday, 18 February 2011

Pace: the final frontier - Q&A

As we enter lower orbit in Pace Wars, your questions answered.

R Hills is God says..
Q: I saw Paul Struthers [BHA Head of Communications] on RUK. He claimed we [British racing] can't afford sectional timings or the weighing of horses. I think its up to you to put the case that we can't afford NOT to have them. 

A: Glenn, I too saw that interview during an excellent broadcast on Racing UK. The BHA clearly haven't a clue how important sectionals are, not least in selling British racing to the US. 

Colin says...

Q: Can I ask you to progress quite slowly? I've always had problems with understanding sectional times.
A: Thanks for hanging tough, Colin. I certainly appreciate your request. I'll do my best to comply. 

Robert says...
Q: Enjoying the blog so far, I'm having a stab at your question.

The reason for the shape of the sectional times at Churchill is mainly the surface. Early speed is king in dirt races because it's hard to quicken, and any lost momentum due to trouble is magnified.
Average US turf sectionals would show a completely different shape; they should be a lot nearer even pace and maybe even skewed towards a faster last two furlongs. 

It is a lot easier to quicken on turf than dirt, and lost momentum doesn't take as much out of a horse; it's more important for a jockey to get his horse relaxed and save energy for the finish. 
Soft ground and synthetic surfaces would change the shape again.

A: There are other reasons, Robert, as I will outline in the next blog. But you absolutely nailed one of the most central. Superbly expressed, too.

horseracingfan says...
Q: Could it be that a horse has to go around two sharp bends and has to slow down to negotiate them, therefore making it difficult to maintain an even pace?

A: Thanks very much for your answer, hrf. It is true that bends do have a significant impact in slowing down horses, particularly in the US. That would not explain why the final, straight two-furlong sectional was the slowest, however. Additionally, there is only bend taken in six-furlong races at Churchill. 

King Cyclops says...
Q: I am not sure about the application of sectional times to British racing. Do you think they would be meaningful here? There are such huge variations in layout, pace and going (even on the sand tracks).

A: Thanks for responding, King. The only problem with sectionals in Britain is getting recorded data. 

With a sufficient sample size, there is no problem applying them to the different conditions you outline. While US tracks are certainly more homogeneous, there are still differences between them for which we need to be allow, caused by variable run-up distances before the timing starts. 
Furthermore, even at a single meeting we need to make allowance for wind changes, track maintenance and rain.

FormPicks says...
Q: What is the point of sectional times if the running rail is moved three yards then an apparent two-furlong furlong bend is now 10 yards less or approx half a second quicker? We even have the situation in the UK where the Oaks is generally run over a longer distance then the Derby!

A: You have answered your own question, FP. To make allowance for different features - such as the rail movement you mention - we only have to make the kind of simple compensation you have worked out. 

Andy Mear says...
Q: Your phrase "energy use" is key. I think that's the fundamental reason why horses are suited to particular tracks. 

Do you believe a horse which has achieved near-perfect sectionals would be capable of doing so in subsequent runs?
Do near-perfect sectionals have as much to do with luck-in-running as ability?
A: Andy, you have brought up two fundamental points which I will be tackling in stand-alone blogs at the end of the primer. 

First, congratulations on your point about track profiling. It isn't exclusively the reason why horses become course specialists, but it has a lot to do with it. 
Cruelly, I am going to make you think some more about your second question, but here are my clues. Forget the narrow idea regarding second run off a layoff, but what might be a general reason for the tendency of some horses to "bounce"? To help, I will give you the phrase "regression to the mean". 
I prefer the term "randomness" to luck, but you are well on the way to the answer.

Thursday, 17 February 2011

Pace: the final frontier III

Thanks very much for all your comments. Such generosity says a lot about your attitude to this great game of ours.

Here is part III of Pace Wars, as my good friend thedarkknight has termed it. We finished the last part by outlining how we account for conditions in assessing times. Now it is time to look at the impact of the track.

Last time, I promised to discuss different surfaces like turf and synthetics. But that has to wait, as I want to develop the ideas in this series at a suitable rate. (Given the title, it would be ironic if I went off too fast.)

* * *


4.2 Surface differences and their implications for pace

The impact of different types of going on race times and sectionals is one thing, but what about the effect of the track's layout and relief?

We don't have to know exactly how much each undulation or bend affects the times which horses run. Such an exercise would be extremely daunting where Epsom racecourse were concerned, for instance.

Instead, we can use a large sample of times over every particular course and distance to infer the combined effect of all features. We know - both from experience and theory - what horses of particular merit should be able to run on a straight, flat, featureless track, and we can compare this expectation with the times they actually do run over each course and distance.

Allowing for error in the mathematical manipulation of this data, the resulting standard (UK and Ireland) or par* (US) time provides an excellent benchmark with which to interpret the time value of future races.

(*Strictly speaking, a par time is more usually applied to the expectation for a particular class level over a given course and distance. It is used more flexibly here.)

4.3 The deeper impact of different tracks

It is not difficult to understand that a track with an uphill finish is more demanding than a flat one. Its standard times will reflect this.

But the rellief and layout of a track is not only evident from the final times of horses, but more particularly from their sectionals. The effects of bends and hills will be instantly apparent from sectional times.

In the same way that we can use data to infer standard times for each course and distance, we can go deeper and develop an expectation of how standard time is optimally achieved there.

This is a more sophisticated task of data analysis. But it is equally within our grasp. And the resulting set of standard sectionals enable us to see far deeper into the equine athlete than many will ever delve.

4.4 Towards the job of track profiling

Let's return to first principles for a second. Remember I referred to the so-called general axiom of pace in the first part of this series? I am now going to express it more formally, at the same time mindful not to become mired in scientific language.

Optimal time is most commonly achieved when sectional times show the least amount of variance.

To rephrase: the most efficient way from A to B is at even tempo. As I stated clearly, however, this is only a general axiom. It does not hold for every horse, every track or every distance. Far from it.

The physical relationship between pace and energy use is fundamental to this axiom. To be rigorous, we might define it as having a perfect application only on a straight track of infinite length and zero undulation. I'll tell you why all three of those considerations are important to the definition in due course.

On such a hypothetical track - and assuming the horse had reached its cruising speed before the timer even starts - the standard sectionals for any given subdivision of the track will be equal to ideal pace for that distance.

(Of course, on the infinitely-long track I stipulated, not only could the horse never reach the end, but its ideal pace in attempting to do so would therefore be infinitely slow, meaning it would remain at the start. This is just one reason why the British Horseracing Authority would never sanction the construction of such a facility - imagine the impact on betting turnover.)

But what about real-life tracks? What can standard sectionals tell us about those? The answer is profound.

4.5 Differentiating between ideal pace and average pace

You will be pleased that we are soon departing from the abstract. Note that many examples during this series are taken from US racing on dirt, only because of the availability of data. I am always mindful that any concepts extracted can be applied universally, and in particular to European turf racing.

Okay, consider the sectional times drawn from a sample of $10,000 claiming races run on the main track at Churchill Downs between 2001 and 2003.

(These are taken from my spreadsheets; you might be able to guess they were barren years for me socially.)

22.2     45.8     1:10.8

These are the average splits that a winner of average ability for the class achieved during that period.

In other words, aided by a running start before the timer began, your common-or-garden $10,000 claiming-race winner would run 22.2 seconds for the first two furlongs, reach the half-mile pole in 45.8 seconds and stop the clock in 1:10.8 after six furlongs.

Let's break those down into sectional times for each individual quarter-mile:

22.2     23.6     25.0

What in the world was going on out there? These aren't anything like equal times; this cannot be ideal pace.

If I showed these splits to someone who had only ever seen British racing and asked them to guess the layout of Churchill Downs, they might infer it had a steep, uphill finish.

Incidentally, if you think it absurd that anyone would be surprised at the layout of a US track, consider the following expression of a presenter on the multi-award winning At The Races channel on Sky 415.

"My first impression is that Belmont Park is extremely flat," said the individual. (Such banality is easy to blurt out on on live television; I have done it numerous times myself. My policy being to avoid personal criticism of others where possible, this is included to highlight the sometimes parochial attitude to racing circuits evident around the world.)

* * *

That's plenty to be getting on with for now. If you are following this series, I am going to leave you with some mental homework. I will not expect it handed in first thing in the morning and there will be absolutely no detention.

What is the reason for the shape of those sectional times from Churchill Downs?

Strictly for your enjoyment, here are three wrong answers:

1) US jockeys cannot judge the pace like "our boys". They go off too fast. British and Irish jockeys are the best in the world.

2) These beat-up, drug-fuelled donkeys can't quicken. Drilling them into the ground is the only way.

3) It is clear that the track is measured incorrectly. The last sectional must be more than two furlongs.

Given you have the motivation to actually read this post, I know you would never have come up with these responses. But they have actually been given to me in another setting, believe it or not.

Best regards and thanks for reading!