Joe Estes' Statistical Analyses

If you don't know who Joe Estes was, then you need to do some reading. Look for the book "The Estes Formula For Breeding Stakes Winners" published by Russell Meerdinck," which is a compilation of articles Joe wrote when he was the editor of The Blood-Horse back in the day (1935-1936). Joseph A. Estes was a meticulous researcher, and he was a numbers guy.  History and pedigree analysts remember him as a real party pooper, because he "poo-pooed" nearly every breeding theory based purely on pedigree and compiled statistics to back up his opinion. Joe's pet peeve was that too many breeders used the names in pedigrees over and above consideration of the individual. He was of the opinion that the individual's performance and conformation were the first and last stop in making any breeding decision. He dismissed the idolization of sirelines and great families, showing that bad runners came from great stock time and time again. He took joy in mocking the disciples of Vuillier's Dosage Theory and Bruce Lowe's Figure System. He felt that inbreeding and nicks were only as good as the parents involved in the mating. Joe was busy quantifying the genetic information through statistics, using the horses' race records - starts, wins, and earnings - to put some sort of qualifier on performance. In his mind, this information didn't fall under the heading of "pedigree," but he was actually adding to our knowledge of pedigrees by helping us understand the difference between average horses and exceptional horses. I'm going to provide a quote from a lecture he gave at the Second Annual Stud Managers Course held at the University of Kentucky in 1952. Joe's talk was entitled "Pedigrees," and it's really good because he makes perfect sense every step of the way. However, here's the last paragraph: "Pedigrees are useful only when we are ignorant of the merit of the individual, and not very useful then. The more we know about the individual and its progeny, the less we need to know about the pedigree. When we have a moderately complete record of the individual and its progeny, the pedigree becomes useless." Joe seemed to be interpreting "pedigree" as just the roll call of the ancestors, and considering many of the lunatic fringe theories that have surfaced over the years based on just that, I can see why he was annoyed. A lot of pedigree geeks are guilty of the same thing today. Those who pay attention to genetic principles, looking at the horses more as livestock instead of aristocrats, consider those hallowed names more as identifying labels on the capsules of what they represent as a genetic package. Progress is made when you consider a pedigree as the genetic blueprint, based on knowledge of the aptitudes and dominance of the horse and its ancestors, and not just knowing how to spell their names. Joe was correct in that once you know what a horse is genetically, that information becomes more important than who its parents were. If a horse is well bred but can't run and can't produce good runners, its superior pedigree is clearly not worth the paper on which it is printed. If the horse has proven that it didn't inherit any of the good stuff from a glittering ancestry, no matter how clever a mating looks on paper, the genetic influences of ancestors buried in that pedigree are most unlikely to come forward. In other words, don't waste your time with proven failures, no matter how enticing the pedigree. Still, Joe's reliance on statistics would have misled him about horses with substandard pedigrees that race at the highest class. Among the so-called Cinderella horses with rags-to-riches pedigrees, successful breeding stallions are hard to find. Silver Charm and Skip Away are classic examples of Cinderella types who went to stud in Kentucky and were effectively stallion disasters. Tiznow could be considered a rare example of a rags-to-riches stallion success story, although his sire was far from a failure and his dam was an exceptional producer in her own right. Estes' numbers can't explain why a superior runner with a low quality pedigree would fail as a stallion, while a lesser-raced but better-bred horse can excel (Malibu Moon being the prime modern example). This key phenomenon makes me a believer in the value of high quality pedigrees, and the value in understanding the difference between a good pedigree and a poor one. Joe Estes' statistical studies showed that the best racemares tend to make the best broodmares, and this true-ism stands the test in every study that has come along ever since. While not all great race fillies become great broodmares, high class fillies do more than their fair share in producing the next generation of high class runners. One of the more lasting of Joe Estes' contributions is a statistical tool known as the "Average Earnings Index" (AEI). The Average Earnings Index compares the earnings of the progeny of any given stallion against the average of all runners contemporary with his progeny. The AEI is best utilized when used in tandem with the Comparable Index (CI), which measures the earnings of the progeny of that stallion's mates when bred to other stallions. Average is 1.00, so any stallion with an AEI over 1.00 is by definition above average. A horse with an AEI of 1.50 is a very good stallion, since his progeny earn 1.5 times the average. This number is put into better perspective if you look at the CI for his mates. A horse with an AEI of 1.50 is doing a good job if his mares' CI is lower than that, say 1.10, so he's moving his mares up. However a horse with the same AEI of 1.50 is a little disappointing if his mares' CI is higher than that, such as 2.00, since it suggests that he's bringing his mares down. This second situation, with an AEI lower than their mates' CI, is not unusual with very young stallions who were at the top of their class and attracted high quality mares. Take, for instance, Bernardini, one of the most impressive young sires in years. The Blood-Horse Stallion Register for 2016 (with statistics good through October 31, 2015) gives Bernardini's AEI as 1.62, which is outstanding, but his mates' CI is 2.88, indicating the very high quality producers that fill his books. Although Bernardini's AEI is less than his mates' CI, no one would suggest that he's a bad sire. So while the two indices are best appreciated when looked at together, it would be misleading to say that any stallion whose AEI is less than his mates' CI was "bringing his mares down." It's all in the context of the situation. That said, any stallion whose AEI is higher than his mates' CI is getting the job done in a positive way, particularly if he's pulling up bad mares (CI under 1.00) to an AEI above average. Love him or hate him, what Joe Estes' statistical studies revealed was that the most reliable way to beat the averages was to use above average sires and dams. In this, Joe Estes was absolutely correct. Copyright by Anne Peters 2016.