Horse running through field

I Want Revenge and the Common Sense of Sire-Line Influence

Roger Lyons on Nicks: Part IV

 In Part II of this series I wrote about the problems encountered by an opportunity-based nick rating system, such as the Blood-Horse Publications system, when dealing with a stallion like Stephen Got Even. I explained why such a system would tend to under-estimate the crosses of a stallion that had dropped in market value, and I used the Stephen Got Even/Mr. Prospector cross as an example. The upshot is that the WTC system rated that cross as an A nick while the competing, opportunity-based system rated it much lower.


Since then, I Want Revenge, by Stephen Got Even and out of Meguial, by Roy (by Fappiano, by Mr. Prospector) (click here to view pedigree) won the Gotham Stakes (G3) by more than eight lengths and became the leading contender for the Kentucky Derby. I don’t mention this just to gloat. Actually, I don’t have to because Steve Haskin, writing for on March 8, 2009 (“Ky. Derby Trail: The Total Package”), quotes I Want Revenge’s owner and breeder David Lanzman thus: “I spent hours on eNicks looking for a mating, and Stephen Got Even is who came up on the screen with the most quality points.” For WTC it’s impossible to buy advertising as good as that from


Besides, the triumph of the Stephen Got Even/Mr. Prospector cross is not the lesson to be taken from the case. True, the Blood-Horse Publications system treats I Want Revenge as an exemplar of the Stephen Got Even/Mr. Prospector cross, but the eNicks system treats it quite differently. This difference is extremely important to anyone who still has doubts about which is the superior system.


Here’s how the Blood-Horse Publications system works as I understand it, and I’ll take the case of I Want Revenge as the subject. It starts with Stephen Got Even and looks to see if he has at least 15 runners from Roy-line mares or stakes winners from at least two of them. It then drops back on the broodmare sire line to see if Stephen Got Even has met either of those conditions with Fappiano-line mares. Then it drops back to Mr. Prospector and does the same search. In this case, Mr. Prospector line meets those conditions, so that becomes the basis for the nick rating calculation.


Now, the Blood-Horse Publications system will go back a maximum distance of three generations, as in this case. If the third-generation sire line doesn’t meet the minimum conditions, then the system drops back to Stephen Got Even’s sire, A.P. Indy, and searches again along the broodmare sire line, counting for crosses to A.P. Indy line. However, the Blood-Horse Publications system doesn’t get that far in this case.


The sire-line search used by that system is entirely determined by the numbers it encounters. It has nothing to do with the identities or influence of the sires involved. Once the minimal quantitative conditions are met, the search stops at that point and calculates the nick rating on that basis. To the Blood-Horse Publications system, I Want Revenge is a Stephen Got Even/Mr. Prospector cross because that cross is the first one to meet the purely quantitative requirements of its search protocol.


Is that an effective way to capture the logic of sire-line influence?


Not according to Andrew Caulfield’s March 10, 2009 TDN column (click here for article) in which his discussion of I Want Revenge focuses on the A.P. Indy/Fappiano cross. Like Andrew, the WTC system knows that Mr. Prospector line is both highly populous and highly diverse, that the Fappiano branch is one of many distinct branches, each with its own peculiarities. It knows that, while some stallions will work well with some of these branches, no stallion will work well with all or perhaps even most of them. It knows, furthermore, that mixing branches together–some of them favorable to a given stallion and others unfavorable–can only serve to wash out the actual effects of individual branches. The WTC system can’t write an intelligent, insightful analysis of the pedigree of I Want Revenge, as Andrew has so ably done, but it knows enough to agree with him about the common sense of the cross.


So, instead of basing a nick rating on all of Mr. Prospector line, it selects the sire-line combination that is most likely to be pertinent to the case at hand and is significantly represented by stakes winners. In this case, it selects the A.P. Indy/Roy combination, the branch of Fappiano line that also yielded Great Hunter, a G1 winner by A.P. Indy’s son, Aptitude. It’s just a matter of common sense.


Common sense is not easy for a nick rating system to come by. Until 1993, the WTC staff looked up nick ratings manually from printouts in very thick binders and made the decisions that are now made by the automated system. When I began automating the WTC method in 1993–turning it into a system–I quickly discovered that my initial sire-line search, organized strictly by the numbers, yielded nick ratings that deviated substantially from those assigned by WTC staff. It was making the same mistakes as the Blood-Horse Publications system because it lacked the intuitive knowledge used by WTC staff to identify the sire-line foundation of a cross. In order to automate the system successfully, I had to discover programmable techniques that would model nothing less than an intuitive knowledge of thoroughbred sire lines.


Common sense has a logic to it, and, once I’d programmed that logic and applied it comprehensively enough to the variety of thoroughbred sire lines, the automated system achieved 100 percent agreement with WTC staff in test samples. The Blood-Horse Publications system, by contrast, went straight to automation, perhaps never had a sufficient level of human input, or was not tested against that input with sufficient rigor. In any event, it operates without the common sense that is so crucial to pedigree evaluation, and it shows.


Besides this important difference in operation, history also accounts for a very broad difference in principle between the two systems. During the first decade of its existence, the WTC system was used exclusively to advise breeders about mating decisions. There was no market for a nick rating system as a medium for stallion promotion. Experience taught that, when a breeder takes an interest in a stallion, there is usually a reason why–perhaps related to the quality of the stallion, its value relative to stud fee, its commercial appeal, its physical suitability for a mare. Breeders were asking the WTC nick rating system for any encouragement it could offer.


It’s not the job of a nick rating system to pick winners and losers, any more than it’s the job of an industry-owned institution like Blood-Horse Publications. That’s a function of the market system. The WTC system does not rest until, within the constraints of common sense, it has exhausted its effort to represent a cross in its most favorable light. That approach serves the interests of breeders as well as it suits the purposes of stallion farms. Make what you will of the fact that the Blood-Horse Publications system has a quite different orientation.

More Means Less for Regional Breeders!

Roger Lyons on Nicks: (Part III)

An elemental consideration in the construction of a nick rating system is the question, what will be taken as evidence of good crosses? For two decades regional breeders have been well served by the WTC system, which has based its nick ratings on unrestricted stakes quality.

From time to time, regional breeders have asked whether or not regionally restricted stakes would provide a more comprehensive basis. It’s also been suggested that a breeder who is trying to breed a good allowance horse might take the winners of open allowance races as the model for how to breed one. It’s a slippery slope.

It sounds sensible to think that evidence of effective crosses should be based on the class of horse you’re trying to breed. Consider, however, the many variables besides the sire-line cross that can determine the performance of a given runner. An exceptional individual can result from a mix of the quality of sire and/or dam, a very good physical match, a favorable inbreeding pattern. Then there is the set of conditions under which a given horse prevails in a given race, including his handling, training regimen, race selection, not to mention racing luck. A horse may be bred from a bad sire-line cross, but perform well at its level for a variety of other reasons. You just can’t tell to what extent a horse’s sire-line cross figures in his performance.

What distinguishes runners of high class–and this is extremely important–is that their performance is much more likely to have been determined by a combination of favorable factors, including the sire-line cross. The highest racing class requires a rich variety of many factors working together. For this reason, runners of high class are much more likely to provide reliable evidence of a favorable cross–or any other value. That’s why, regardless of the class of horse you’re trying to breed, you should look to horses of the highest class for your breeding model.

This logic has been turned on its head at the Blood-Horse Publications website dedicated to its nick rating. They confuse the standard of evidence by saying that “patterns of successful breeding turn up in good and bad runners.” It’s not a matter of where a breeding method can “turn up.” In order to see that a bad runner is bred from a good cross, you must already know it’s a good cross. You can’t make that determination on the basis of bad, mediocre, fairly good, or just lucky runners. It has to be based on truly superior runners.

What counts as evidence should be determined by a balance between two opposing requirements of a nick rating system: 1) the selection of winners must be broad enough to encompass the array of sire-lines that might be crossed in the population; 2) the system must take as evidence winners of the highest possible class. In order to make a system useful to everyone, it has to yield numbers large enough to evaluate the range of sire-line crosses that occur. However, the farther down the class scale it goes, the more compromised that evidence will be.

Graded/group stakes provide the most reliable basis for evidence of the quality of sire-line crosses. The problem is that winners of those stakes do not adequately represent the array of sire-line crosses that must be covered by a comprehensive nick rating system. So, the WTC system includes unrestricted stakes. For 20 years that has served to strike a balance between the need for reliable evidence of good crosses and the numbers that are needed.

If restricted stakes are not needed, then, considering how they would affect reliability, what compensating value could they possibly add? Since the Blood-Horse Publications nick rating system is the only one that includes restricted stakes, let’s see what they have to say.

Surprisingly, Blood-Horse Publications has little to say about it, other than that competing products “are not able to accurately identify and report specific regionally-successful nicks.” I suspect that the designers of the Blood-Horse Publications system have not reflected on the impact a large volume of restricted stakes would have on the reliability of those so-called “regionally-successful nicks.”

I suspect, too, that they haven’t thought through the idea of “regionally-successful nicks” beyond its potential as a selling point. Had they done so, they would surely have discovered the disastrous consequences that idea has for a comprehensive sire-line nick rating system. More about that in my next post.