Nature versus Nurture: Are Champion RaceHorses Born Or Made?

Dec 27 2007 Published by under Brain & Behavior, Genetics, Horses, Journal Club

tags: , , , , , ,

Seattle Slew (1974-2002), the only undefeated horse to win the Triple Crown (1977).

I have been thinking about a paper that was published last week, that analyzed the effects of "nature versus nurture" on the development of a champion racehorse. In short, this paper found that the effects of a horse's pedigree is minor when compared to its environment .. the combined effects of training, diet, choice of races entered, jockey skill and of course, injuries, which are unpredictable. In short, the exorbitant stud fees paid to breed a mare to a prized stallion are not an honest signal of the stallion's genetic quality. However, upon reading this paper, I think the situation is more complicated than how the authors present it.

Evolutionary ecologists Alastair Wilson and Andrew Rambaut at the University of Edinburgh in Scotland compared the stallion nomination fees for 554 stallions currently (or recently) active at stud with the winnings and lifetime earnings of 4,476 of each stallion's male and female ancestors that were foaled in the United Kingdom or America between 1922 and 2003. They found that environmental factors such as training, diet, strategic race entry and the skill of a horse's jockey accounted for 91.5% of the variation in a horse's winnings.
"It seems much more likely that people who can afford to pay high stud fees can also afford to manage and train their horses well," observed Wilson.
According to Wilson and Rambaut's data analysis (figure 1, below), if the predicted progeny lifetime earnings returned some portion of the breeder's investment in the sire's stud fee, the line would slope upwards -- a line with a slope of 1 means that the offspring lifetime earnings return equalled the stallion nomination fee. However, this line is flat -- actually, if you look closely, you will notice that it has a positive slope of 0.02, which corresponds to a 2-cent payback for every dollar spent for stud fees;

However, while it is true that the offspring of expensive stallions tended to win more over their lifetime, Wilson stated, genes played only a small role. In fact, the authors found that 8.5% of a racehorse's success was due to genetics. While a genetic effect of 8.5% sounds trivial to the casual observer, genetics typically accounts for only 1-2% of survival rates in the wild where environmental conditions vary tremendously and where an individual's survival can depend on luck as much as anything else. But because the authors deemed "success" in horse racing as being heavily dependent upon just one factor, winning races, they found that genetics plays a much bigger role in this situation.
"There are good genes out there to be bought but they don't necessarily come with the highest price tag," Wilson pointed out.
Thus, the authors determined that a stallion's stud fees are not an honest signal of that horse's genetic quality if the breeder's objective is to maximize lifetime prize winnings.
Even though this conclusion seems simple enough, the authors neglect to mention quite a few other factors that play a part in breeders' stallion choices. For example; the amount for a newly retired stallion's stud fees are determined by the group of people who purchase lifetime shares in that stallion, and this price is dependent upon the horse's purchase price, and that, in turn, is based on his performance on the racetrack as well as his lifetime earnings and his pedigree. Thus, after a stallion has been siring offspring for five or six years, his stud fee will be adjusted so it reflects his overall quality as a sire of race horses, because several crops of his progeny have provided breeders with examples of their racing abilities. In view of this information, I'll bet the authors will find a much tighter statistical relationship between a stallion's nomination fees and offspring winnings if they limited their dataset to those stallions who have been at stud for at least seven years.
Another consideration is that at least a few stallions have been recognized for their genetic contributions to their daughters, moreso than to their sons. Even though fillies and mares do not typically run in races that have large prizes, their wins are an important indicator of their competitiveness and abilities. Further, mares also provide valuable genes to their offspring so, like stallions, their primary success also lies in the breeding shed. These are examples of several types of less easily analyzed information that a breeder uses when designing pairing between racehorses.
So basically, it is true that a breeder might get what he pays for -- and less -- when breeding his mares, it is more likely that breeders are using several criteria to make their stallion choices other than relying solely on potential progeny lifetime winnings.
This study was published last week in the Royal Society journal Biology Letters.
Alastair J. Wilson and Andrew Rambaut. 2007. Breeding racehorses: what price good genes? Biology Letters (doi:10.1098/rsbl.2007.0588). [PDF]
BBCNews (quotes).

6 responses so far

  • Josh Rosenau says:

    There's also an issue that the adaptations which contribute to fast racing also produce structurally weaker horses, so horses which could be good earners are more likely to wind up injured and taken out of competition. There's a tradeoff there, and jockey quality figures into how likely a horse is to be in a position to be injured. If the paper models injury as a purely random process, it'll miss real factors.

  • you're correct, josh. in fact, many breeders will specifically note when the horse has retired sound, without injury, since a fair number of race horses are run until they have an injury, at which point, most of them are retired.

  • David Harmon says:

    Also, (1) the stud price could be held down by market forces, and (2) the competitive environment isn't static either. Given that human athletes certainly show an overall trend toward better performances over time, I wouldn't be surprised if racehorses did the same.

  • It could just be that the horses at the top are so close, even the slightest difference is valued. Even if paying extra for the best genes gives you only a horse that can run 2% faster... that could turn a second-place into first.

  • Lynda says:

    Racehorse stud fees are based on the combined history of racing success of a stallion and that of his progeny. Although research may show that only 8% of a racehorse's success is based on his genetic make-up, thoroughbred breeders know how indicative a horse's bloodlines are of his racing future. Bloodlines not only determine the conformation of a horse, they help determine the horse's heart. As with humans, attitude is everything, so with the Thoroughbred.

  • Glenn Craven says:

    I'm getting in on this topic two and a half years late, but with some valuable (I believe) comments.

    Last things first; Lynda's comment on the basis of stud fees ignores an incredibly significant factor: Commercial appeal. A stallion's racing success does figure into the equation. So does the performance of his "get," i.e., progeny. But being from the "right" auction-attractive bloodlines and getting the occasional huge winner -- rather than many, many good to very good horses -- creates a wide discrepancy in stud fees.

    Elusive Quality, sire of Smarty Jones, Quality Road and others, stands for $75,000 and once stood for $100,000. Elusive Quality won nine of 20 starts, but only two of those races were even Grade 3 in status (that is, two grades below the Kentucky Derby and the other "biggest" races). He has a propensity for getting one or two "home run" horses out of each crop. Bust most of his foals are average at best. At this writing, just 66 percent of his foals of racing age have made at least one start, and only 45 percent are winners. (That includes 2-year-olds who haven't had a full chance yet.) Elusive Quality's average earnings per starter from his foals are $67,729. That shows in his progeny's median earnings of $22,850.

    Meanwhile, Slew City Slew (sire of the great Lava Man, but few others of which you might have heard), while being a son of Triple Crown winner Seattle Slew (which should mean something, genetically) stands for only $3,500. He was a far better racehorse than Elusive Quality, winning 11 of 42 starts for $1,166,296, including two Grade 1 races, the Oaklawn Handicap and Gulfstream Park Handicap. His earnings are 2.5 those of Elusive Quality. ... He's also a more consistent stallion, getting 80 percent runners from all foals, 59 percent winners, average earnings of $60,021 (which are only slightly less than Elusive Quality's) and median earnings of $22,235 which are virtually the same.

    Why the difference in stud fee? ... Slew City Slew's average weanling or yearling sells for around $13,000 at auction; his 2-year-olds about $20,000. Elusive Quality's weanlings sell for $35,000, his yearlings for $80,000 and his 2-year-olds for nearly $86,000 on average. (Which still are massive losses; the market is finally catching on that EQ has always been overpriced.)

    So it can't be said with authority that stud fees are entirely -- or even to the most rational degree -- based on racing success of the stallion and his progeny.

    As for this study pegging nurture over nature in determining which horses perform best, there certainly must be truth (it stands to reason) that those who an afford to breed expensively, can also afford to raise and train that horse with the best resources money can buy. When I like a certain horse at the auction and he is bought for small money by a small operator, it often makes me cringe, because he's probably going to a cheap track to be treated like a cheap horse.

    But even if only 8 percent of the horse's success is based on genetic makeup, it cannot be ignored that 8 percent -- at top levels of competition -- is *huge.*

    Dr. Fager set a world record for one mile on dirt, carrying a staggering 134 pounds (he was a heavily handicapped favorite) over the Arlington Park track in 1:32 1/5 seconds to win the 1968 Washington Park Handicap.

    Even trained by the same person in exactly the same way (or with the same care and individual attention), a horse just 2 percent less physically gifted than Dr. Fager would have run the race 2 percent slower -- in 1:34 2/5 -- and been beaten by about 10 lengths. A horse 8 percent less-blessed genetically might have run 1:39 3/5 and lost by 35 or more lengths ... further behind than the nearest finishers in Secretariat's colossal Belmont Stakes victory.

    When you're up against the best, being 92 or 98 percent as good isn't nearly enough.