The Myth of a Phalaris/Non-Phalaris "Mega-Nick"
(This is the text of an article originally written in 1995)

There has been considerable discussion in recent years about the so-called Phalaris/Non-Phalaris (P/NP) "mega-nick". The concept has troubled us from the outset because the evidence presented in support of the theory was static in nature, relying on data frozen in time. For example, when the theory was offered by Rommy Faversham in 1991, the question was asked, "Why is it, then, that the Phalaris line has never come to dominate broodmare sirelines?". The data showed that between 1964 and 1990, NP sires dominated the leading broodmare sire lists in terms of champion broodmare sire, AEI, and progeny earnings. We'd like to offer contradictory evidence suggesting that this is not really the case. Using North American graded stakes winners as the source, we note that for GSWs born prior to 1980, 35.4% represented the P broodmare sire line and 64.6% the NP broodmare sire line. However, for GSWs born between 1980 and 1987, the figures are 49.3% P and 50.7% NP. For GSWs born after 1987, they are 63.5% P and 36.5% NP. Clearly we are in a phase where P broodmare sires have expanded dramatically in the population of GSWs to the point where they far exceed the influence of the NP broodmare sires. And for all intents and purposes, GSWs are an indication of future breeding trends.

The original study also focused on the success of P/NP runners in the highest class races such as the Kentucky Derby, Belmont Stakes, Breeders' Cup Classic, and the Breeders' Cup turf races. There truly is a high percentage of P/NP horse that win these races. The question is whether this success symbolizes a true affinity of bloodlines or whether there are other factors at play. Two issues come to mind. First is distance, and second is the European influence. The Derby, Belmont, Breeders' Cup Classic, and Breeders' Cup Turf are run at 10-12f where stamina is a key ingredient. For the most part (and obviously there are exceptions), P sires represent speed in pedigrees (through sires like Raise a Native and Nasrullah) while NP sires represent endurance (through sires like Ribot and Princequillo). The blending of these aptitudes enhances the probability of a pedigree balanced in speed and stamina, historically a key component of classic race competitiveness. The Breeders' Cup Mile (as is the Breeders' Cup Turf) is populated with European horses. Our data show that the dams of European G1SWs since 1973 are only 33.1% P and 66.9% NP. Therefore, there is a higher probability of an NP dam among Breeders' Cup winners coming from Europe. Admittedly this is anecdotal evidence, but so is the evidence from the original study when it comes to these particular races because it suggests no cause and effect relationship. There could be many explanations apart from a "mega-nick" to explain the observations made. We have presented just a couple. we will later come back to the very large impact of NP mares on European racing and show that, despite this, there is no indication of P/NP supremacy even in an environment where NP mares are pervasive.

The original P/NP study also alluded to the success of P/NP sires relative to P/P sires. This allows us to introduce methodology to demonstrate the statistical basis for a nick. We will illustrate with hypothetical data. Let's say that among a population there are 80% P sires and 20% NP sires, as well as 60% P broodmare sires and 40% NP broodmare sires. That's essentially the raw distribution of bloodlines. If there is no advantage to any particular cross of bloodlines, the appearance of each cross should reflect the overall distribution. For the data cited, there should be 48% P/P; i.e., 60% (P broodmare sires) of 80% (P sires). Similarly, there should be 32% P/NP (40% of 80%), 12% NP/P (60% of 20%), and 8% NP/NP (40% of 20%). The total does indeed come out to 100% and is the predicted distribution of the particular crosses given the existing distribution of bloodlines in the population. Deviations from the prediction will suggest a true affinity or nick.

The real data, using North American GSWs, is as follows:

For GSWs born before 1980 the distributions are 48.4% P sires, 51.6% NP sires, 35.4% P broodmare sires, and 64.6% NP broodmare sires. The distributions for the various crosses are shown as %actual and %predicted, followed by the ratio of %actual to %predicted.

For GSWs born between 1980 and 1987 the distributions are 67.9% P sires, 32.1% NP sires, 49.3% P broodmare sires, and 50.7% NP broodmare sires. The distributions for the various crosses are:

For GSWs born after 1987 the distributions are 79.8% P sires, 20.2% NP sires, 63.5% P broodmare sires, and 36.5% NP broodmare sires. The distributions for the various crosses are : P/P 50.1/50.7, 0.99; P/NP 29.7/29.1, 1.02; NP/P 13.4/12.8, 1.04; NP/NP 6.8/7.4, 0.92.

The data tell us that before 1980 there was truly an advantage to P/NP and a disadvantage to P/P. We ascribe this to the speed on speed characteristics of the early P/P crosses, mainly through Nasrullah and Raise a Native. In succeeding time frames we see not only an explosive growth of P sires and P broodmare sires in the population of top class runners, but a dramatic narrowing of the gap in advantage for P/NP. At this point, there is little to support any difference at all unless one claims that ratios of 0.99 and 1.02 are measurable. Most important, however, is the trend. Shifts in bloodlines are dynamic, not static. The only way to get a handle on what's happening is over time. The trend is clearly in favor of P/P, probably because of the emergence of stamina-oriented P broodmare sires who bring balance into the P/P pedigrees. These include P broodmare sires such as Nijinsky II, Alydar, Roberto, and Buckpasser.

Our foreign racing data is only for G1SWs in Europe since 1973. Here, even though there are significantly more NP broodmare sires, the data show a true advantage for the P/P cross and, surprisingly, the NP/NP cross. The general distributions are 66.6% P sires, 33.4% NP sires, 33.1% P broodmare sires, and 66.9% NP broodmare sires. The distributions for the various crosses are (again, %actual/%predicted and ratio of actual to predicted): P/P 23.6/22.1, 1.07; P/NP 43.0/44.5, 0.97; NP/P 9.5/11.1, 0.86; NP/NP 23.9/22.3, 1.07.

The data for North American sires of GSWs since 1973 follows.

The general distributions (i.e., of the sires themselves, not their graded stakes-winning progeny) are 65.4% P sires, 34.6% NP sires, 34.9% P broodmare sires, and 65.1% NP broodmare sires. The distributions for the various crosses suggest no real advantages, except perhaps for NP/P sires. A look at a more recent timeframe might be useful.

Finally, we believe these studies provide good evidence that there is no statistical basis for the claim of a P/NP "mega-nick". If anything, the trends indicate a growing importance for P/P in North America. There already seems to be a P/P advantage in Europe. We can't really predict the future of bloodlines, but if the current scenario plays out, we wouldn't be surprised if someday we talk about Phalaris the way we now talk about Eclipse, and the hot topic of conversation will be the Nearco/Non-Nearco nick.