Too Many A Nicks? I Don’t Think So.
Roger Lyons on Nicks: Part I
The TOBA’s Blood-Horse Publications used to run ads for Werk Thoroughbred Consultants, Inc.’s nick rating and its other pedigree consulting products. Blood-Horse Publications stopped doing that late last year because WTC became their competitor. Well, that’s not exactly true. It’s not that WTC started a publishing empire. Rather, Blood-Horse Publications became a competitor of WTC by entering the pedigree evaluation business. In support of that new venture, they have staged an all-out assault on a product and company from which for many years they happily banked advertising revenue.
I want to set that straight because, from what they’re saying at a Blood-Horse Publications website dedicated to their new nick rating, you’d think the WTC nick rating is the one that was born yesterday. Of course, everyone knows that’s not true, so you can only get it by implication. It’s quite vivid, however, in a particularly thoughtless, disrespectful, and idiosyncratic post, dated October 29, 2008, titled "No Apologies: TrueNicks Reports the Bad with the Good," and written by Byron Rogers, Australia’s William Inglis & Sons’ National Sales Manager and a partner in the Blood-Horse Publications nick rating system.
The central point of that post is that their nick rating sets an uncompromising standard of measurement despite the interest of its stallion sponsors in associating a stallion with the broadest possible range of A nicks. The post was occasioned by a complaint from a stallion manager about the small number of A nicks, relative to the number yielded by the eNicks system of WTC. True only to form, their post defends its rating by attacking the WTC rating.
Thoughtless: The post exclaims in bold-face type that their "competition (WTC, presumably) routinely awards nearly 60% of the entire population an ‘A’ or better rating!" The post explains that this proportion was yielded by "a series of trials of up to 1,000 horses." Its reference to "the entire population" implies that they used samples that were representative of all registered foals. I submit that one or the other of the two claims they make–one of them explicit, the other implicit–is quite simply false. Should either of them fail, then the statement itself is wholly false.
Of course, you don’t need to survey samples of 1,000 horses in order to find out the proportion of A nicks. Samples of 100 will yield very consistent results across samples, but only if the samples are consistently representative of the population of interest–in this case, "the entire population." If you select samples from different sectors of the population–say, from the most elite runners, the major commercial population, a regional auction catalogue, a racing card, and the foal population at large, respectively–then, in that case, you’ll get five significantly different frequencies of A nicks.
So, I requested from WTC a breakdown of nick ratings for the 136 horses entered on a Saturday card at Turfway Park. Again, as any statistician will tell you, that sample size is quite sufficient. It consisted of 39 percent A nicks or above, including 6.6 percent A+ nicks and 6.6 percent A++ nicks. Keep in mind that these are horses that made it to the races, a vast majority of them winners. The frequency of A nicks would be significantly smaller for a group truly representative of "the entire population."
If their sample came up with 60 percent A nicks, then, clearly, they sampled a group that is utterly unrepresentative of the population, without so much as disclosing the source of their sample. Thoughtless.
Disrespectful: That same post also charges, without any historical evidence whatever, that the WTC nick rating was "created with the sole intention of ‘selling seasons’," a wildly irresponsible indictment of the very origin of the WTC nick rating.
That indictment forgets that the WTC rating, unlike their rating, was created 20 years ago, long before it had entered anyone’s head that a nick rating could be used to sell seasons, long before they created their nick rating. The WTC nick rating could not possibly have been "created" with the "intention of selling seasons," much less the "sole intention." Its potential for revenue lay entirely in its appeal as a measure used for advising breeders as to stallion selection.
Nevertheless, they publish a "study" amateurishly concocted to misrepresent WTC’s nick rating, and then, on that basis, launch insults at WTC, its products, the stallion farms that subscribe to WTC’s service, and the intelligence of the post’s readers. Disrespectful.
Idiosyncratic: A curiously cryptic banner ad recently appeared at The Blood-Horse website: "If everything is an A++… is it really an A++?" I’m guessing this bizarre ad is intended to reinforce the cynical, utterly misleading, and, yes, idiosyncratic message of Byron Rogers’ post. Here’s a story that explains just how odd that message is.
When WTC created its nick rating 20 years ago, there were no letter grades. There was just the variant score, and it was used strictly for its input into WTC’s advice to breeders. As soon as the nick rating was available for that purpose, clients began to ask what it meant. So, after a time, WTC applied the letter grades–A, B, C, D, F–scaled to the variant.
Anticipating that clients would only want to breed A nicks, WTC scaled the grades in such a way as to reserve broad latitude in recommending stallions–25 percent A, 25 percent B, 25 percent C, and 25 percent D and F ratings. To assume a normal distribution, relegating A nicks to only a small area of the curve–say, 13 percent–would only have served unduly to exclude from consideration creditable stallions representing reasonable potential for a given mare.
Subsequently, when clients began requesting individual nick ratings, WTC added the A+ and A++ grades. A+ nicks are as different from A nicks as A nicks are from B nicks, and the A++ category is reserved for a very small proportion of crosses. Contrary to what Blood-Horse Publications would have you believe, the WTC scale is well adapted to distinguishing the relative value of sire-line crosses.
It’s not as if Blood-Horse Publications, by some marvel of statistical ingenuity, discovered that 13 percent of all actual crosses are A nicks. No, they decided that only 13 percent would be A nicks. They decided the question of grade distribution–just as WTC had done–when they set up their grading scale. Whoever sets up a grading scale is the decider of how many crosses will be given an A, at least initially.
By the simple logic of cause and effect, the practical use of the WTC system was bound to affect the actual distribution of grades. Over time, the proportion of the population qualifying as A nicks increased because breeders were using nick ratings. It would be irresponsible to raise the standard down the road in order to limit the number of A nicks just because more foals are bred from good crosses. After all, to reflect good crosses is precisely the role of an A nick rating.
Here’s my point. When the deciders at Blood-Horse Publications set up their grading scale, they assumed that WTC’s nick rating was born yesterday, just like theirs was. They disregarded the years during which it was used to advise WTC clients, the ten years when it was distributed in batches by major regional and some Kentucky stallion farms by way of WTC’s Software for Stallions product, and the five years when it was distributed wholesale by the hundreds of thousands through WTC’s eNicks website.
It’s preposterous to suppose that, after all that, the proportion of A nicks bred in the population would decline, rather than increase. When I write that there are more A nicks being bred now than ever before, Byron Rogers calls it "a dangerous attempt to make a virtue out of a vice." Nonsense. The truth is that, when people buy a product, it’s because they intend to use it. When Blood-Horse Publications and its partners feel endangered by the truth, it’s time for them to establish a closer relationship with it.
Even if the proportion of A nicks in the population has increased to as much as 35 percent, which is probably not the case, that is hardly a surprising result of 20 years during which nick ratings have been available and, especially during the last decade, used in the industry on a grand scale. In view of that, the actual population-wide increase in A nicks is really quite small.
The reason is that, although nick ratings are used very intensely in the major commercial tier, for obvious reasons, A nicks are far less frequent among horses bred in contexts where expectations of commercial potential or of a racing career are more remote. After all, one-third of the foal crop never start.
In spite of the demand that grew around WTC’s nick rating, a demand that Blood-Horse Publications is eager to exploit, they assume that its immense popularity had no effect whatever on breeding decisions. Idiosyncratic.
None would question how fitting it is for an organization owned by the Thoroughbred Owners and Breeders Association to defend standards. However, the manner in which Blood-Horse Publications pretends to do so–at least in regard to its nick rating–falls far below the standard that ought to be expected of an industry-owned institution. They would have done well to copy WTC’s approach, which I have closely monitored over the years. Its products and marketing have always been done thoughtfully, respectfully, and in accord with a community of interests.
In his next post (look for it next week), Roger Lyons examines the dubious practice of using quantitative estimates of "opportunity" in the statistical evaluation of sire-line crosses.
Roger Lyons is President and CEO of Roger Lyons Consulting, Inc., specializing in pedigree information services and pedigree consultation. He has participated in the development and distribution of the Werk Nick Rating since 1992.