Over the last decade, nicking gurus have done a fabulous job of purveying their trumped up science to the masses. Nick ratings for hypothetical foals are available on an increasing number of farm websites and sellers in private transactions frequently brag about the nick rating of the horse they’re attempting to sell.
In previous pieces, we’ve itemized the numerous shortcomings for readers, but simultaneously mentioned that there is some degree of usefulness in the nicking sciences if weighted accordingly. If the resourceful breeder has covered the essential components of a successful mating i.e. conformation, racing aptitude/style, etc, and two stallions appear to be on equal terms, then perhaps a nick rating can be used as a tiebreaker.
But since then, we’ve uncovered the nail in the coffin… the nail that has finally rendered modern nicking theory to its final resting place next to other inherently flawed breeding theories such as dosage and birth rank. This ‘final straw’ forces us to throw nicking off the breeding theory map altogether. But before we divulge the reason for eulogizing nicking, let’s first revisit some of the more well-established shortcomings that many breeders continue to ignore as they rush to the breeding shed with the excitement of a favorable nick rating.
Lack of Transparency
Currently, mare owners who request a nick rating are provided with a five cross pedigree of the hypothetical foal along with the Roman Dosage Profile, a quality rating, quality points, total points, the actual rating as a function of the standard A to F academic scale, and a listing of any inbreeding found within five generations.
Collectively, the numbers appear to represent a form of high science that will aid breeders in producing better foals. The problem lies with the company’s unwillingness to disclose most, if not all factors used to derive the various ratings.
The most obvious lack of disclosure is the sample size used to formulate the final letter grade. Mare owners have no idea if the rating was compiled using a sample size of 1, 10, or a 100. Given the company’s affinity for throwing numbers at mare owners, you’d think they would want to list the number of horses used for a particular rating… unless, of course that number reveals a science so badly flawed that ratings are often based on a statistically insignificant sample.
Referring back to the principle of using nicking as a tiebreaker, imagine a mare owner has narrowed his search to two stallions. In an attempt to tip the scales, the mare owner goes to the stallion’s respective farm’s websites and pulls a nick rating for each hypothetical cross. The mare owner finds that when his mare is crossed with stallion X, the resulting nick is an A++, while the cross with stallion Y results in an A, convincing the mare owner to book his mare to stallion X.
But what if the mare owner had information about the sample size used for each rating and learned that the A++ rating had been formulated by using three runners, while the A rating was based on dozens of runners? The mare owner in this situation would have finalized his decision based on ratings being purported as equally derived. The simple solution would be for the company to provide the sample size to it’s customers, but when simple and obvious solutions are being ignored, there’s usually a shortcoming of some sort being concealed.
Take for example, the alleged nick between daughters of Lyphard with A.P. Indy and his sons. Assigned a letter grade of B+, a breeder with a daughter of Lyphard may be quick to assume that the rating’s purveyor went to elaborate lengths to come up with such a neatly-packaged indicator. The truth is, daughters of Lyphard have only produced seven starters from A.P. Indy and/or his sons. To make a broad-reaching inference based on just seven starters is an enormous leap of statistical faith.
With such concise letter-grade descriptions, the whole concept of nicking is geared primarily for providing simple, easy-to-understand answers that do little to reveal methodologies or the complexities of planning a successful mating. This certainly works to the benefit of those with a financial stake in nicking, as it is always more efficient and profitable to sell a simple, neatly-packaged concept as opposed to selling a complex methodology that requires more of an explanation.
The main problem here is that we’re dealing with a very dynamic question: What is involved in a successful mating? Nicking theory tries to answer this question with an all too simple letter grade, which is terribly insufficient to answer such a complex question.
Lack of Replication in Lower Percentiles
Nicking theorists are often quick to point out how their theory applies to the upper levels of horse racing i.e. unrestricted and graded stakes winners, but they never mention whether or not the same phenomenon exists in the lower percentiles of racing. After all, if this were the case, and we saw the same dispersal of nick ratings in the bottom tiers, the theory would have no practical use. John Gaines saw this and commissioned his own study of lower level claimers to see if they paralleled the populations used by nicking pundits:
“The big issue is, that while there are all of these statistics, these pedigree nerds, the people that are promulgating all of this scientific nonsense, there is never any control group. It is all selective, self-serving, worked-over information. I did a study one time with Dewey Steele about nicks, and we found the worst horses, those finishing last at Thistledown and Charles Town, had the same nicks as the horses that were winning the classics. One of the dumbest things that’s done is that they take an infinitesimal sample of maybe three or four or five horses and pompously say that is a nick. From my scientific point of view, a sample that is that minuscule is worthless.”
Failure to Distinguish Generations Removed
A handful of stallion farms in central Kentucky now allow prospective clients to obtain nick ratings for hypothetical matings to their stallions. This may have been a mistake on the part of the nicking company as breeders are now able to compare and contrast how the ratings change as a function of different variables. Even the layperson will notice irregularities that don’t adhere to scientific principles.
As we began to tinker with a few of these sites, one thing stood out immediately: the rating did not change as a function of the number of generations a certain sire was removed from the hypothetical foal. For example, if you wanted to see how your daughter of Afleet nicked with the influence of A.P. Indy, the rating would remain the same if you’re breeding directly to A.P. Indy or to one of his sons. Common sense tells us that our expectations should not be identical for a direct mating with A.P. Indy versus a mating with Stephen Got Even. Yet, nicking theory in its current format fails to distinguish the difference.
Omission of the Female Influence
Perhaps the most obvious flaw in nicking theory is the omission of the female influence. If nicking pundits are able to point to certain statistical probabilities based on the sire line, they should be able to point to similar probabilities based on certain female families and blue hen type mares.
Imagine a five generation pedigree. In cases where the nick rating is based on a paternal influence 3 generations removed and a maternal influence 3 generations removed, the final rating fails to incorporate over two-thirds of the entire pedigree. No theory out there can claim to take 100% of all factors into consideration, but nicking theory falls well short of being an inclusive theory.
Less Research by Mare Owners
An indirect result of the nicking craze is the effect it has on the behavior of mare owners. Too many breeders (and yearling buyers for that matter) are allowing this over-simplified method of planning matings to take up the bulk of their planning process. Many will simply make a phone call to get the latest ratings before booking their mares. Less and less time is being spent on researching progeny results, conformation, market trends, etc. Where as this should be viewed as a tie-breaker at best, it is all too often being used as a starting and ending point.
Again, just to reiterate… prior to discovering this latest shortcoming, we gave nicking the credit we felt it deserved, as a tiebreaker. But with this new information (new to us anyway), it needs to be thrown out altogether, allowing breeders to spend time with more useful and statistically valid approaches. The deal-breaker:
Omission of Opportunity
In describing the presence of a phenomenon (or lack thereof), the foundation has to be built first upon opportunity. To develop a descriptive statistic or rating that ignores opportunity is futile at best, but that’s precisely what the nicking gurus have been doing all along.
Common sense tells us that in order to evaluate the success or failure of crossing a broodmare sire line with a particular sire line, the first piece of necessary information is opportunity, or how many times has this already been tried in the past? Though there may be dozens or even hundreds of anecdotal stories attesting to the strength of a cross, unless we examine these anecdotes as a function of overall opportunity, the resulting information is unequivocally useless, right? Wrong, according to the nicking pundits.
The current methodology for deriving nick ratings, as described by Roger Lyons is described with the following equation:
Total SW’s bred from a certain cross Total SW’s from the broodmare sire line
Total SW’s from the sire line Total SW’s from the entire population
Now admittedly, I am anything but a math expert. But from my days as a psychology major where I was required to take multiple classes on statistics and research methods, I’m certain I recall professors emphasizing the importance of measuring overall opportunity before analyzing subsets. Without some idea of how large the entire population is, there is no reason to move on to analyzing the dynamics and statistical importance of subgroups. No where in the equation above will you find an indicator of total number of live foals and/or starters representing the cross.
Nicking pundits argue that because the entire sample representing a sire/broodmare sire cross is diluted with inferior lines of foundations sires such as Mr. Prospector or Northern Dancer, the analysis should only encompass the stakes winning population. Not only does this defy all forms of logic and common sense, but it also ignores the fact that by ignoring non-stakes winning progeny, you ignore those progeny hailing from high-end strains who fail at the track. Certainly, a tabulation of these foals can contribute to a discussion of whether or not a cross is successful or not. Current nicking models actively avoid this crucial data that could help breeders make better decisions.
My guess is that nicking companies see the obvious flaw in their methods, but turn a blind eye as a function of practicality. Even though they have an elaborate database of five generation pedigrees, it is likely cost-prohibitive to acquire race records for the hundreds of thousands of horses in their system. So instead of presenting their customers with a complete picture of the dynamics behind their ratings, corners are cut in the form of studying only stakes winning horses.
Over the years, I’ve tolerated the nicking sciences. While bearing in mind the shortcomings mentioned above, I’ve even referred to the ratings on occasion, but for the most part, I’ve utilized them strictly as a means to get a complimentary five-generation pedigree. But after this most recent revelation surrounding the omission of opportunity, it’s time to lay the nicking era to rest and give it it’s due as a badly flawed and misleading pseudoscience that likely set the development of pedigree theory back an entire generation.