The Derby Points System and the Decline of the American 2yo: A Statistical Analysis

Post Reply
Tessablue
Posts: 3800
Joined: Fri Sep 13, 2013 11:29 am
Location: Boston

Mon Mar 25, 2019 10:36 pm

There's been a lot of talk lately about the state of the modern thoroughbred and the potential downstream effects of altered early development and workload on young thoroughbred bones. Bisphosphonates have made quite a bit of news, and just today a new practice of sales testing for these drugs was implemented.
 
Another factor which we've discussed a little on this forum, but which hasn't picked up a lot of momentum nationally, is the issue of 2yo workload. Although studies of breakdowns have found that a lack of 2yo starts is a risk factor for injury (sources https://www.paulickreport.com/horse-car ... -involved/ and https://www.governor.ny.gov/sites/gover ... Report.pdf, links kindly provided by Sparrow Castle) 2yo racing is often only talked about in a negative context. "Babies are forced to race" is a common emotional appeal by anti-racing groups, and it's not unusual to find similar sentiments in the comments section of certain racing websites, especially after fatal breakdowns. However, from a biological perspective it is very important that young bones are exposed to load and exercise so that they can grow stronger and adapt to the biomechanical forces involved in running.
 
So, why does the title mention the Derby points system? This system was first instituted in 2013 with a partial objective of keeping precocious 2yos who underperform at 3 out of the race. Juvenile stakes races are worth extremely little in this system, and it took several years for the Breeders' Cup Juvenile itself to be worth more than 10 points to the winner. So this begs the question: has 2yo racing been devalued in this country? After all, what race drives the public consciousness of the sport more than the Derby?
 
I started out by looking at Derby runners, which are by far my largest and most accessible dataset. I averaged the number of 2yo starts per runner from 1990-2018. Note that this applies exclusively to horses who raced in America at 2; European juveniles in particular tend to run more often, and for these purposes I'd like to focus solely on American racing.
 
Charting the average number of 2yo starts per runner looks like this, with the line delineating the debut of the points system in 2013:
Image 

Although moderately cyclical, these results at least visually appear to demonstrate a drop in 2yo races starting in 2013. Appearances can be deceiving, so I ran a simple statistical test between two groups: the fields from 2007-2012, and the fields from 2013-2018. I restricted the years here because horses in general are making fewer starts and I wanted to control for that effect by using only the years immediately preceding the points system.
 
The p-value from this test was 0.003, indicating a very significant difference between the two groups. Although it's impossible to derive causation from this test, it does indicate that something has caused our Derby horses to reduce their starts at 2.
 
So, is this because they are racing less in general, or are they starting later as well? After all, if only 3yo preps are worth considerable points, why bother aiming for 2yo stakes races as well? To this end, I charted the median debut month for each Derby field. Mercifully, it wasn't quite as boring and slow as it sounds!
 
The results are shown below. The number on the y-axis refers to the median debut month, so 9 = September, etc:
Image

As you can probably see, the difference here is extraordinary. The statistical test seems a bit cursory, but just for the record, the p-value is 0.0008. Starting immediately in 2013, Derby horses began to make their 2yo debut much later, if at all. This fits in with a general devaluing of 2yo races, and also represents a troubling trend away from the early workload that is necessary for developing strong bones and adapted joints.
 
Now, you might be thinking to yourself: "Tessa, you idiot, the points system was supposed to do this. Obviously, if you specifically prevent precocious horses from entering the Derby, you end up with fields that aren't as biased towards 2yo racing."
 
To this I would say: you're right! So, to check whether this trend is limited only to Derby horses, and conscious of the fact that the Derby drives everything in this sport, I looked at BC Juvenile fields as well. If the Derby is the only field impacted by this system, it stands to reason that our best 2yos would still be racing at the same rate as before, right?
 
As before, I averaged the number of starts per American entrant. Note that this information does not extend as far back, because it is much harder to find BC pps (for the 80's races I literally just found videos of the post parades on youtube and wrote the records down). I will continue to add to this, but for the moment, I think enough information is present to make a statistical judgment. Here the line is in 2012, because that was the first 2yo crop to be impacted by the points system:
Image
 
The above figure is less obvious than the Derby one, and as you can see, the average number of starts dropped in 2010 and 2011 in advance of the points system in 2012. However, the average number of starts has not rebounded to pre-2010 levels, and a statistical test of the years from 2005-2011 vs. 2012-2018 results in a p-value of 0.007. In other words, for whatever reason, our top 2yos are making fewer starts than they used to. Last year, Instagrand was shelved after two starts in anticipation for his Derby run. How often, either consciously or unconsciously, has this strategy impacted juvenile careers in the past several years?

Presently I am trying to look at Juvenile Fillies fields to see if the trend is present among them as well (currently, it looks like it isn't), but I just don’t have as much access to those fields and going through Equibase is too time-consuming. But I'll update with those findings when and if they happen, and if it turns out our fillies are making roughly the same number of starts as usual, it may be an even stronger indicator of the power of the Derby points system.
 
It remains to be seen whether this year's crop will follow these trends, but some discussion questions to consider in light of this information:
 
-Do you think the data presented are convincing RE: a link between the Derby points system and a decline in 2yo starts? What information do you think might help make this connection stronger or more convincing?
 
-Are you concerned by this information? Do you expect to see wider unanticipated effects from the points system?
 
-Do you think there is a way to promote both a points system and 2yo racing in America? How might we better strike a balance between the two?
 
Thanks for reading!
Last edited by Tessablue on Sun Mar 31, 2019 3:22 pm, edited 1 time in total.
User avatar
Sparrow Castle
Posts: 6020
Joined: Wed Sep 18, 2013 6:44 pm

Tue Mar 26, 2019 12:19 am

This is amazing information! But, as usual with data, my mind jumps around with many additional questions. I'd love to convene some trainer focus groups to learn about any shift in their philosophy towards training possible future champions pre and post point system.

I wish I had more time to test this out, but family issues are taking over right now. I just looked at Instagrand's workouts. He had his first one at 2f on 04/15/2018. They are pretty regular around his two 2018 races, on 6/29/2018 and 8/11/2018. He had two more works after that race, on 08/23/2018 and 09/10/2018. His next work was 12/30/2018, then nine more works before his 3/9/2019 race. He's had two works since the Gotham Stakes.

So it looks like Instagrand was out of training for three months and a few weeks over the Fall of his two year old year, after what might be a fairly normal start time for his training/racing career? Galilean, also trained by Hollendorfer, had his first work 06/29/2018. Game Winner, Champion 2 yo and BC Juvenile winner (Baffert) had his first 2f work on 04/21/2018, trained and raced through the BC, then back in training 01/09/2019 and regularly since, not significant time out of training.

My biggest question around the Derby contenders making fewer starts in this point system era has been are they starting training later in their two year old years or are their races spaced out more with multiple works in between. It seems to me that regular workouts would contribute to bone-building at least almost as much as actual races.

There might too many other variables to when a horse is first put in training and whether or not they take breaks to make much sense of it, but dang I'd love to see it. I would love to see how this pattern compares to other Derby contenders past and current, but recognize how impossible that would be.

I just wish the Jockey Club would open up their Equibase data, as well as their equine injury database, so those inclined could happily play around with the data. We actually might come up with the ideal training and racing program regarding long-term health and safety for maybe the majority or at least some horses. ;)
User avatar
Miss Gladiator
Posts: 10
Joined: Fri Mar 15, 2019 11:59 am

Tue Mar 26, 2019 1:02 pm

Sparrow Castle wrote:
Tue Mar 26, 2019 12:19 am
My biggest question around the Derby contenders making fewer starts in this point system era has been are they starting training later in their two year old years or are their races spaced out more with multiple works in between. It seems to me that regular workouts would contribute to bone-building at least almost as much as actual races.
It would be interesting to have data about that. Workouts tend to be run slower and shorter than races, so I'm not sure they would have the same kind of impact. How much of a strong workout and strong gallops each week would a horse need to get the equivalent of a fast mile and a sixteenth race, for instance?

The impact the Derby points system has had on the sport in a few short years is pretty mind-boggling. Less and less racing at two, less sprinting, and favorites winning the world's most chaotic race every single year... I have to say, though, it's at least reasonably fun, even with less than inspiring crops. The Road to the Kentucky Oaks, on the other hand... does anyone care about which fillies get the points?
BaroqueAgain1
Posts: 10189
Joined: Sat Sep 14, 2013 6:16 pm

Tue Mar 26, 2019 1:35 pm

The Road to the Kentucky Oaks, on the other hand... does anyone care about which fillies get the points?

After Serengetti Empress's poor showing, IMO everyone is chasing 'Yesterday. ;) :D
User avatar
Miss Gladiator
Posts: 10
Joined: Fri Mar 15, 2019 11:59 am

Tue Mar 26, 2019 1:45 pm

BaroqueAgain1 wrote:
Tue Mar 26, 2019 1:35 pm
The Road to the Kentucky Oaks, on the other hand... does anyone care about which fillies get the points?

After Serengetti Empress's poor showing, IMO everyone is chasing 'Yesterday. ;) :D
:D
By the way, just to be clear, I wasn't saying that I don't care about the fillies. In fact, this year I care about them a great deal more than about the colts; they seem like a good bunch, and at least I know most of them won't retire prematurely to go to stud, and many will certainly race at 4 or 5. But I can't remember a time when a really good filly wasn't in the Oaks because of a lack of points.
aethervox
Posts: 196
Joined: Wed Sep 18, 2013 11:48 am

Tue Mar 26, 2019 3:36 pm

I have past performances for the BC Juv Fillies from 2010-2018, if that helps; I also have a list of all the runners of the Juv Fillies, since it started.

I also have PPs for the Oaks (not sure how far back on those).

PM me with an e-mail and I'll send them to you.

I sent you a PM about other data I have, but I understand if you would like to see the original numbers, not my transcriptions.

aethervox
User avatar
Gemini
Posts: 93
Joined: Thu Jan 10, 2019 9:26 pm

Sat Mar 30, 2019 3:52 pm

This is a great read, but I have to admit some of it went over my head. I took statistics in college and immediately forgot most of it. Why is a p-value of 0.0003 significant?
Tessablue
Posts: 3800
Joined: Fri Sep 13, 2013 11:29 am
Location: Boston

Sun Mar 31, 2019 3:21 pm

Gemini wrote:
Sat Mar 30, 2019 3:52 pm
This is a great read, but I have to admit some of it went over my head. I took statistics in college and immediately forgot most of it. Why is a p-value of 0.0003 significant?
Thanks, and sorry, I should have explained that better! The p-value essentially tells you the probability that the differences between groups are due entirely to chance. So a p-value of 0.05 basically means that there is a 5% chance you have a false positive and the groups only seem different because of some coincidental variation. Meanwhile, a p-value of 0.0003 means there's only a 0.03% chance that this is the case, so we can pretty conclusively state that the groups are different. 0.05 is considered the standard in research, although there's a well-deserved push to get rid of the p-value as a measurement of statistical significance... that's a whole other story though!

I think the question of workouts vs. races is a fantastic one, and I wish there was an easier way to get at it. From a biomechanical perspective, brief but high-intensity works should still be sufficient to remodel and strengthen bone (in labs that study biomechanical adaptation, cyclical bone loading once per week is more than enough to induce bone growth). However, I have no idea if it would work to the same degree as racing, and there's also the cardiovascular side to consider as well.

I also wonder if those quick works at 2yo sales are actually good for them in the long-term (as long as they aren't on those awful bisphosphonates). This is kind of a morbid question, but does anyone recall any fatal breakdowns during those sales? Hypothetically, they should be very rare.

And yes, the road to the Oaks is considerably less dramatic than the road to the Derby. Does anyone recall any fillies actually being kept out by the system? It doesn't really seem like it comes up often in the prep season.

(and thank you aethervox, check your pm box! :) )
User avatar
Sparrow Castle
Posts: 6020
Joined: Wed Sep 18, 2013 6:44 pm

Sun Mar 31, 2019 5:06 pm

Tessablue wrote:
Sun Mar 31, 2019 3:21 pm
Gemini wrote:
Sat Mar 30, 2019 3:52 pm
This is a great read, but I have to admit some of it went over my head. I took statistics in college and immediately forgot most of it. Why is a p-value of 0.0003 significant?
Thanks, and sorry, I should have explained that better! The p-value essentially tells you the probability that the differences between groups are due entirely to chance. So a p-value of 0.05 basically means that there is a 5% chance you have a false positive and the groups only seem different because of some coincidental variation. Meanwhile, a p-value of 0.0003 means there's only a 0.03% chance that this is the case, so we can pretty conclusively state that the groups are different. 0.05 is considered the standard in research, although there's a well-deserved push to get rid of the p-value as a measurement of statistical significance... that's a whole other story though!

I think the question of workouts vs. races is a fantastic one, and I wish there was an easier way to get at it. From a biomechanical perspective, brief but high-intensity works should still be sufficient to remodel and strengthen bone (in labs that study biomechanical adaptation, cyclical bone loading once per week is more than enough to induce bone growth). However, I have no idea if it would work to the same degree as racing, and there's also the cardiovascular side to consider as well.

I also wonder if those quick works at 2yo sales are actually good for them in the long-term (as long as they aren't on those awful bisphosphonates). This is kind of a morbid question, but does anyone recall any fatal breakdowns during those sales? Hypothetically, they should be very rare.

And yes, the road to the Oaks is considerably less dramatic than the road to the Derby. Does anyone recall any fillies actually being kept out by the system? It doesn't really seem like it comes up often in the prep season.

(and thank you aethervox, check your pm box! :) )
Re the workouts, Equibase only records the official (AKA timed) workouts. In my experience, horses go out for long gallops too that aren't recorded. That's especially so when they are first put in training or coming off layoffs, but also in between timed workouts. I just don't see any way we could capture this for specific horses in a data-set.

Yes, there have been breakdowns during the two-year old in training breezes. I don't think it's common though. Don't remember which sale, but I don't think it was more than a year or two ago, that I saw one that had to be euthanized. I don't like those quick breezes at all. Think about all the training those quick babies must have had before the breeze at the sales. I think more and more breeders and consignors are opting to not go for the quickest time with their horses, and I hope that trend continues. It seems to me prospective buyers should be able to get all the info they need just by watching how the horses move on the track. It doesn't have to be at breakneck speed.
User avatar
Sparrow Castle
Posts: 6020
Joined: Wed Sep 18, 2013 6:44 pm

Sun Mar 31, 2019 5:16 pm

For whatever it's worth...

Travis Stone Verified account @TravisStone
6h6 hours ago
You read about it often but the points system, which started in '13, has not "slowed down" the pace in the @KentuckyDerby. In fact, its been faster based on fractions:

Last 6 years (13-18): 46.3 - 1:10.9
Prior 6 years (07-12): 46.8 - 1:11.3
Prior 12 years (01-12): 46.4 - 1:11.0

https://twitter.com/TravisStone/status/ ... 1280159748
stark
Posts: 4953
Joined: Thu Oct 03, 2013 9:55 am
Location: SoCal

Sun Mar 31, 2019 5:24 pm

Sparrow Castle wrote:
Sun Mar 31, 2019 5:06 pm

Yes, there have been breakdowns during the two-year old in training breezes. I don't think it's common though. Don't remember which sale, but I don't think it was more than a year or two ago, that I saw one that had to be euthanized. I don't like those quick breezes at all. Think about all the training those quick babies must have had before the breeze at the sales. I think more and more breeders and consignors are opting to not go for the quickest time with their horses, and I hope that trend continues. It seems to me prospective buyers should be able to get all the info they need just by watching how the horses move on the track. It doesn't have to be at breakneck speed.
Do they use whips during those exhibitions of speed?
I've found it easier to tear up tickets at 8/1 instead of 8/5.
User avatar
Sparrow Castle
Posts: 6020
Joined: Wed Sep 18, 2013 6:44 pm

Sun Mar 31, 2019 5:34 pm

stark wrote:
Sun Mar 31, 2019 5:24 pm
Sparrow Castle wrote:
Sun Mar 31, 2019 5:06 pm

Yes, there have been breakdowns during the two-year old in training breezes. I don't think it's common though. Don't remember which sale, but I don't think it was more than a year or two ago, that I saw one that had to be euthanized. I don't like those quick breezes at all. Think about all the training those quick babies must have had before the breeze at the sales. I think more and more breeders and consignors are opting to not go for the quickest time with their horses, and I hope that trend continues. It seems to me prospective buyers should be able to get all the info they need just by watching how the horses move on the track. It doesn't have to be at breakneck speed.
Do they use whips during those exhibitions of speed?
They carry whips. Have a look at the videos of the most recent under tack sale...
http://www.fasigtipton.com/2019/The-Gulfstream-Sale#/
Tessablue
Posts: 3800
Joined: Fri Sep 13, 2013 11:29 am
Location: Boston

Sun Mar 31, 2019 8:45 pm

SC, I wish I could find you an article I read a few weeks back about a recent 2yo sale. It featured a buyer talking about how fast times are overrated and sub-10s don't mean what they used to. I also don't see much point in pushing 2yos to run as fast as possible, but I do wonder if the timing of the training in any way provides some benefits. Wish I knew more about the topic.

The Derby pace is a topic I feel pretty good about, though- I think it's a bit dangerous to rely on pace averages for that information, because we've seen a lot of variability in the Derby track in recent years. As an example, the pace in 2016 looks brutal on paper but occurred over a highway, whereas 2010 doesn't look that fast on paper but was in fact a complete meltdown over a sticky track.

I've actually been tracking the Derby pace for a while by looking at raw numbers as well as the spread of the field and the positioning of the top finishers (were they in front, way behind, etc.). I haven't found a way to quantify it yet, but it's not too hard to use the eye-test to designate races as slow, medium, fast, and meltdown.

Before the points system was instituted, we had a fairly unusual spate of meltdowns. The races in 2012, 2010, 2007, 2005, and 2001 could all qualify. However, from a statistical standpoint they were balanced out by slow and medium years- 2011, 2009, 2008, 2006, 2003, and especially 2002 were all slow to moderate.

Since the Derby points system was instituted, we've had one meltdown: 2013, which was under highly unusual circumstances. We haven't had one since, but we have had one slow pace (2015), one slow-to-moderate one (2014), and now a string of fast-but-not-insane ones. So although the averages may indicate that the pace hasn't changed, what we're seeing is basically a tightening of pace scenarios- and the insane meltdowns of earlier years have become comparatively much rarer.
User avatar
Sparrow Castle
Posts: 6020
Joined: Wed Sep 18, 2013 6:44 pm

Sun Mar 31, 2019 9:59 pm

I'd love to read that article if you come across it again, Tessablue. I do hope buyers are rethinking the issue of sub-10s in those breezes. I'm not saying don't get the babies ready for the under-tack breezes, just don't push the horses to the edge of the cliff. They can start training at the same time, just at a more reasonable pace progression. These babies shouldn't have to run in record time to break their maidens.

I think we've all heard the stories about the home-breds whose owners were told by the folks at the breaking/training farms that they have a real fast one on their hands and send him/her to a good trainer. That's part of their jobs to get a sense for the good ones. Some know even by just watching them race the other babies in the paddocks. I've heard that at least some of the babies who get the speed records at the sales are turned out for a time after the sales because they're worn out. Seems to me, if you have to delay racing them what have you really gained by buying a speed record breaker for that $1mil or more?

And I agree with you about the Derby pace tweet. It's that old trap you fall in when you aggregate aggregates, lol. That's why I added the "for whatever it's worth". Someone mentioned in that Twitter thread that a big difference since the point system is that it affects the staying power of the speed horses that earn their way in, and now the jocks can't ignore those horses as cheap speed, or something like that.
Post Reply