Dosage: Pedigree & Performance
The Concept of Dual Qualifers
The fusion of Dosage and the Dual Qualifier (DQ) concept never was meant to be a magical formula for betting on the Kentucky Derby. Unfortunately, thanks to a largely uninformed turf media, that is how most of the public perceives it. In an industry more concerned with style and self-promotion than in substance and intelligent inquiry, it should come as no surprise that many turf writers, exemplified by superficial thinkers like Andrew Beyer, Steven Crist, Bill Oppenheim and others, continually misrepresent the meaning and purpose of Dosage and DQs. Whether out of ignorance or malicious intent, that is the reality. Instead of constructive debate, these publicly irresponsible and intellectually dishonest individuals use ridicule and accusation to make a case against Dosage. The methods they use and the illogic of their arguments raise questions about their own intellectual capabilities. Despite the constant barrage of insults, these misguided souls have never been able to challenge the data or the accuracy of the Dosage model which has been confirmed through the years by numerous investigators and researchers with far more credible credentials. Apparently they can't think in abstract terms due either to under-education or to a profound lack of academic motivation. It is quite possible that their personal commercial agenda plays a role as well. Regardless of the reason, these attacks do nothing to undermine the validity of the basic model we have proposed. Instead, they expose the weaknesses and failures of the attackers.
From the beginning, the DQ idea was offered as a model of the American classic horse. That model featured a set of individual characteristics which together differentiated classic horses from their contemporaries. Based on historical observation it became evident that the vast majority of American classic horses could be defined in terms that didn't apply to almost all other Thoroughbreds. We identified three unique characteristics that made these horses different, the latter two being closely inter-related. First, they were bred to stay a mile and a quarter as young three-year-olds. Second, they had matured earlier than average and, third, they were among the elite runners of their generation as two-year-olds, exhibiting exceptional class virtually from the beginning of their racing career. The rationale behind this hypothesis is that under the unique conditions of classic racing these factors are critical for success. For one, we felt that Grade 1-level pace pressure requires genetic suitability to the distance. Furthermore, that suitability will be fully expressed under the race conditions. In addition, we believed that early maturity combined with extraordinary class is needed to handle the unique configuration of classic racing, particularly in the Kentucky Derby. The Derby is like no other race. The Derby field is uniformly of exceptional quality with superstars from every region of the US and, occasionally, other countries converging to confront one another for the first time. Most of the Derby starters will never again compete against a field with such depth of talent. The pace of the Derby is almost always very quick, putting special emphasis on strength and physical development, as does the usually large field size where those same attributes help the runners to overcome the inevitable crowding and bumping that occurs and which can be intimidating to those less mature. Advanced physical development is obviously an asset once we realize that the challenge of ten-furlongs comes at a point in time when the starters are still very young, the equivalent of teenagers at best. Finally, maturity helps the horse to maintain a relaxed demeanor in the midst of the Derby's circus-like atmosphere and spectacle, enabling the horse to stay focused on the task at hand. Thus it is easy to rationalize the characteristics described as essential for classic performance.
Having proposed the hypothesis, the next logical step was to develop a methodology for testing it. As mentioned, Dosage figures are a good first approximation of distance potential and the Dosage Index (DI) was chosen as the metaphor for genetic suitability to ten furlongs. This was done for the sake of simplicity even though we had established earlier that the Center of Distribution (CD) was the more accurate statistical tool and that the Dosage Profile (DP) was the most useful tool for pedigree interpretation. For convenience, a high ranking on the Experimental Free Handicap (EFH) was selected as an approximation of early maturity combined with extraordinary ability since the EFH is a rank order of the year's best two-year-olds. It is quite possible, in fact it is virtually certain there are other, perhaps better ways of evaluating the qualities implicit in a DQ. The DI does not capture emerging aptitudinal prepotence in young sires and the EFH is, to a large extent, a subjective exercise. Of course, since the primary goal of Dosage research is to understand the long-term aptitudinal evolution of the Thoroughbred, the list of aptitudinally prepotent sires will always to some extent lag behind current developments in breeding. Similarly, the EFH selection process is continually undergoing change, in part because those who do the selections change and they will apply their own values to the process. Nevertheless, DI and EFH ranking were a reasonable alternative, they could be tested easily and they were readily understandable to the average person interested in Thoroughbred racing and breeding. The term Dual Qualifier was assigned to any horse meeting both the DI and EFH criteria. The intention was to demonstrate, if possible, a positive correlation between DQ characteristics and classic racing success.
Historical research at the time (the early 1980s) revealed that no Kentucky Derby winner since 1929 had a DI over 4.00 despite the fact that about one-fourth to one-third of all stakes winners did. That suggests we should have seen as many as ten to fifteen Derby winners exceeding the DI 4.00 figure over the previous half century. Surprisingly there were none. The observation is consistent with lower DIs correlating with a more likely distance pedigree. Similarly, a review of the EFH over the previous decade revealed that every Derby winner was ranked within ten pounds of the highweight, confirming both early maturity and very high class. With this information in hand we set about to test the hypothesis into the future which we have done to this point in time. The results generally confirm the hypothesis that DQ characteristics do indeed define the American classic horse. Between 1973 and 2006 there have been 93 American classic races in which at least one DQ started. Of these, 53 (57.6%) were won by a DQ despite only 19.4% of the fields being comprised of DQs. The result is an Impact Value (IV) of 2.97 where IV is a measure of success vs opportunity. In this case, only 19.4% of starters possessed the characteristics of a DQ yet they won 57.6% of the races. In a random world, where no trait or set of traits represents an advantage, the IV would be 1.00. Therefore, DQs, by virtue of their distance pedigree and high-class early maturity alone, win almost three times as many classic races as we might expect.
Now it is true that DQs in the Derby over the last decade have not fared as well as they had in the past, with only one DQ winners since 1997. On the other hand, DQs in the Preakness and Belmont have performed as well or better than they had previously (see here). The failure of DQs in the Derby in contrast to their continued success in the last two legs of the Triple Crown is difficult to understand and the subject of speculation. That the public is generally unaware of this information is directly attributable to the media's focus on Dosage only in the context of the Derby.
One possible way to explain the difficulty experienced in the Derby by DQs in recent years is the concept of DI inflation. Since 1940, the average DI of Derby winners has increased by about 0.04 DI units per year. A consequence of this trend is that by the year 2020 it is projected that half of all Derby winners will have a DI exceeding the historical 4.00 guideline. The Thoroughbred is changing. It is continuously being infused with more speed. There is pressure from owners and trainers to shorten the distance of major races as fewer horses remain competitive over a distance of ground. Under these conditions it should come as no surprise that the DI criterion alone may be insufficient for determining whether a pedigree meets the Derby standard. In simpler terms, the standard is changing. At the same time, the continued success of DQs in the Preakness and Belmont suggests that, unlike the Derby, there is a more stringent selection process at work. Everyone wants to take a shot at Louisville and in huge fields a dream trip by a horse otherwise ill-suited to the challenge will often determine the outcome. Perhaps it is not merely coincidence that between 1991 and 2005 the trend line of the Beyer Speed Figures of Derby winners is in a downward direction while the trend line of the DI of Derby winners is in an upward direction.
This brings us to the critical question of whether it is time to reevaluate the DQ concept. That pedigree, early maturity and class are the cornerstone of classic performance is not the issue. The discussion instead revolves around the best ways of assessing DQ qualities in classic contenders.
Dosage has always depended on a sire's confirmed demonstration of prepotence for aptitudinal type. Quite often it takes years to establish such prepotence in a statistically significant manner. Yet many classic starters are by young sires that have not yet had the opportunity to exhibit aptitudinal prepotence. In those cases there may be a complementary methodology for assessing pedigree. While using raw Dosage figures as a beginning point, we might consider notionally modifying them based on a sire's known winning distance profile. Alternatively we can use, for example, Tomlinson distance ratings. Regardless of how we do it, the objective is to develop a technique that increases the accuracy of the pedigree analysis. For example, 2005 Derby winner Giacomo is by Holy Bull and out of a Stop the Music mare. Although he has a DI of 4.33 which is just marginally above the DI 4.00 guideline, his Tomlinson Distance Rating is well over 300 indicating suitability to a classic distance. With the Tomlinson data in hand and knowing that he was very close on DI anyway, one might have looked at his pedigree differently. Combining this knowledge with the fact that Giacomo was the 7th highest rated two-year-old of 2004, assigned 122 pounds on the EFH (four below the highweight), we probably should reconsider whether he did, in fact, have the attributes of a DQ. At 50-1 odds certainly it was worth some thought.
It should also be noted that the EFH itself has undergone profound changes over the years, changes that question the utility of the "within ten pounds of the highweight" guideline. When Ferdinand won the Derby in 1986 as a DQ rated at 116 pounds on the EFH, he was tied for 36th best among the two-year-olds of 1985. In 2006, Derby winner Barbaro was tied for the 18th best two-year-old of 2005 on the EFH even though he was assigned only 114 pounds and failed to qualify as a DQ by just two pounds. Similarly, Bluegrass Cat, second in the 2006 Derby was tied for 14th at 115 pounds, one pound below the cutoff. Now either the folks who generate the EFH are using a different standard than they once did or they are telling us that today's two-year-olds just don't measure up to those of the past. I doubt seriously whether that is their intent. However, reality suggests that "within ten pounds of the highweight" alone may no longer be as useful as it was and could be complemented with, for example, "among the top 20 or 25" two-year-olds. Considering that Barbaro and Bluegrass Cat appear to fit the traditional mold of the classic American horse, this modification may be justified.
Another weakness of the EFH is the set of rules required for inclusion. Two-year-olds that did not finish among the top four in open stakes races are ineligible for consideration. That seems a bit harsh considering the EFH is meant to identify the best of the best in terms of talent. Certainly the name of a race doesn't necessarily define the quality of its entries. Take the case of 2004 Derby winner Funny Cide, a state-bred that never raced in open stakes competition at two yet earned the 4th best two-year-old Beyer Speed Figure of 2003. His performances against open company as an early three-year-old easily confirmed his elite status as a juvenile, and with a DI well below 4.00, one could reasonably have argued for his DQ status. The same is true in the case of 2004 Derby winner Smarty Jones, who also did not race in open stakes competition but earned the highest two-year-old Beyer Speed Figure of 2004 and had a DI well below 4.00. Now I am not arguing in favor of dropping the EFH in favor of Beyer Speed Figures. I am suggesting complementary ways of assessing two-year-old form. They could be Beyer Speed Figures, my own Performance Figures, BRIS Speed Ratings or even Thoroughbred Times Performance Rates. In the end, the objective is a better understanding of the horse's qualifications as a classic candidate. We certainly should not be wedded to an inflexible formula originally intended solely as a means of testing a hypothesis.
We can go back even further in time to the late 1980s to see other examples of Derby winners excluded from inclusion on the EFH by the rules but which clearly possessed DQ traits. One is 1988 Derby winner Winning Colors which raced only twice as a two-year-old. Significantly, she handily defeated the eventual two-year-old filly champion Epitome in a maiden race at Saratoga and, following an injury, convincingly won an allowance race late in the year. That she continued to progress without missing a step into her three-year-old campaign is more than enough to indicate that had she not been injured, she surely would have been among the very best two-year-olds of the year, and with a DI inside the historical guideline.
Sunday Silence, which won the Derby in 1989 with a qualifying DI, is still another example. After breaking his maiden at six furlongs by ten lengths in the spectacular time of 1:09 and 2/5ths, he followed in late December with a head loss at equal weights to the colt that went on to become the winter book favorite for the Derby. Sunday Silence also progressed impressively into his three-year-old campaign and had shown he was essentially on equal terms with the colt that early on was considered most likely to take the roses on the first Saturday in May.
Both Winning Colors and Sunday Silence were extraordinarily talented juveniles and well within the DI guidelines. It is a virtual certainty that had they been given the opportunity to compete during the year in open stakes company they would have achieved a very high EFH ranking.
Another example is 1996 Derby winner and qualifying DI colt Grindstone which was weighted very low on the EFH on the basis of just two starts and who suffered a season-ending injury in his stakes debut. How good was he? He broke his maiden by five lengths in early June at Belmont Park earning a PF of -29. No other North American two-year-old in 1995 earned a superior figure in a maiden race until October. Grindstone took up where he left off as a young three-year-old and moved forward. Based on his debut performance and sustained quality at three, we can be confident that had he not been injured he would have been among the elite of his juvenile season.
Barbaro, Giacomo, Smarty Jones, Funny Cide, Grindstone, Sunday Silence, Winning Colors - seven Derby winners since 1988 with pedigree and two-year-old performance characteristics arguably consistent with the traditional DQ model but excluded from that designation because of self-imposed rules that fail to take into account all of the information that is available. Imagine the success rate of DQs in the Derby since 1988 had other criteria been employed in identifying them. But even with these exceptions, the DI/EFH criteria have more than adequately demonstrated and confirmed the validity of the DQ model and defined the American classic horse.
If we remember that it is the principle of the DQ and not the preservation of an archaic methodology that matters, a more liberal approach makes a lot of sense.
Updated, Spring 2008