Grading the Elite Thoroughbred
For those of you who read this blog, you would be aware that I am not a great fan of the current graded stakes system as a framework for selection for genetic merit. There are a couple of reasons for this, but the main reason is the variability of form in stakes races around the world on a year by year basis make selecting for merit based on this framework, especially in lower class stakes races, next to worthless. I believe I am not alone in this assessment.
Back in 2011, as a project for Dan Rosenberg as advisor to Sagamore Farm, Professor Robert Losey of the University of Louisville completed a very good study on the broodmare selection as it related to performance and production. You can read the entire report here, which I suggest you do if you have the time. Among his findings Losey found that Listed winning mares were only marginally better producers of stakes winners than any other categories - allowance, other winners, and unraced mares. The difference in each case was not statistically significant and unraced mares exhibited only slightly lower strike rates than listed SWs, allowance winners, and the other winners category in producing stakes winners. What was damning in terms of the value of Listed Stakes winning mares is that the commercial market paid significantly more for Listed winners, even though their subsequent production didn't warrant it. Not only was it a bad bet buying a Listed stakes winning mare but it was a bad bet buying one out of it over a similar yearling out of an allowance winning or unraced mare.
Unfortunately due to the size of his sample, Losey was unable to separate out Gr III winning mares, which we believe share a similar trait or fate to those Listed winning mares. They are of limited genetic value. In his regular column in the Thoroughbred Daily News, Bill Oppenheim has often noted that his "A Runner" is the best measure he can find of genetic merit. Bill has frequently commented that this type of horse is generally a Gr. I or Gr. II winner. While I disagree with the way that he comes up with his identification of elite performance (prize money is a terrible metric to use as it is lowly heritable and not allocated to class correctly which is why his APEX ratings have Canadian stallions up high in them), I thoroughly agree with his definition of class - Gr. I and Gr. II winners are in effect all that counts for the breed, the rest is just noise.
In our own studies we have found that in genetic terms, Listed and Gr III winners are generally speaking no different to allowance/handicap level horses. That is, they generally share the same variants with little genetic variability between these groups. There have been a few published studies by other genetic companies that have used Listed and Gr. III winners as "elite". We think that this is a terrible mistake and a good reason for finding spurious associations in these studies. The reason for this we believe is that "trainer effect" is high in these lower level stakes races that skew the results to make average horses better than they actually are. To explain this effect, a horse with one trainer can be a good handicap/allowance level horse while the same horse with another (and his or her associated vet and jockey) can win at Listed or Group III level. This is more apparent in races at two and three, where they are racing against their own age, and also races where they are running against their own sex and age where you can get Listed or Gr. III status almost by turning up on the day to beat a bunch of what subsequently turn out to be very mediocre horses. Peeling back the genetic layer and looking at the variants associated with class it is apparent that while there are many ways to be an elite horse, the chasm between a genuine Gr. I/Gr. II horse to a Gr.III/Listed level horse is large.
I should bold the word genuine in that last paragraph because it is important. Every year we see horses that achieve Grade/Group One status that don't genuinely deserve it, even in what are supposed to be the best of the best races (George Vancouver anyone?). The form going into some of the races that are afforded Gr. I status is suspect, but it is even worse coming out of it with many of the runners never winning races approaching Gr. I class. Its a shame Graded/Group Committee's around the world can't give the status out retrospectively. Part of me would like to see the graded/group stakes system abolished and replaced with a ranking system, as it isn't helping the average breeder select properly for genetic merit, but the other part of me tells me to leave it alone as it represents a significant market misallocation which we can take advantage of.
In a roundabout way, maybe the American Graded Stakes Committee has come to a similar conclusion. On Aug. 12 they announced that for 2014 they will raise the minimum purse required for a stakes race to be eligible for grade I ($300,000) or grade II ($200,000) status. In making the changes the comment was made....."There had been no raise in the purse minimum for graded stakes for several years and the committee felt it was important to have higher purses for our top tier races," said Dr. J. David Richardson, chairman of the AGSC, which is part of the Thoroughbred Owners and Breeders Association. "Many non-stakes races now have purses comparable to graded races and we believe our best races should be set apart from the rest. With the declining foal crops we may see a movement toward somewhat fewer but more valuable graded races."
While the AGSC plan to raise the minimum for Gr.III races to $150,000 in 2015, based on what we have seen, if the AGSC is interested in using the stakes system as a framework for genetic merit, not just for prettying up the catalog page, I wouldn't bother. Concentrating on making sure that Gr.I and Gr. II races genuinely represent the best of the breed (like say having 10 minimum starters to be awarded the status; consistent drug testing protocols; natural progression of races; reduction of competing similar events) will do more for the breed as a whole than making sure Gr.III races are run at a prize money standard. Genetic progress is going to be made by properly defining the elite, like Oppenheim does with his "A Runners", not adding more noise to the process as Losey's study has shown.
That said, I do applaud the AGSC for making this change, especially as the foal crop sizes decline in North America. Their decision is in contrast to the Australian Pattern Committee who earlier this year decided to upgrade 41 (yes 41!) races to group/graded status. That decision can't be seen to be anything but a fattening of the catalog page for breeders, with no regard to genetic merit at all.