Dual Qualifiers in Perspective

It is generally a given that subtlety is not a particular strength of the popular press.  In order to attract the widest readership possible, the strategy is...simple.  Exactly!  Make it SIMPLE for the reader to understand.  Say it all in the lead paragraph of the story because relatively few people are interested in the details.

When Dosage hit the racing world in the early 1980s through the good graces of Daily Racing Form breeding columnist Leon Rasmussen, the press ran with it, exploited it and made it SIMPLE.  It was that mentality which led to the widespread and erroneous idea that a horse with a DI over 4.00 couldn't win the Kentucky Derby.  What we actually wrote at the time, and what appeared in Daily Racing Form, was that research had discovered no Derby winner with a DI over 4.00 as far back as 1940 even though a sizeable percentage of the Thoroughbred racing population exceeded that figure.  Other researchers later extended the observation back to 1929.  How "can't win" had evolved from "hadn't won" is still a mystery, although the former is certainly more dramatic and, in a sense, threatening.  Converting an academic observation into a challenge to traditional thinking apparently creates reader interest.  Thus began the incarnation of Dosage as a "magic formula".  Never mind that the original publication of the research emphasized Dosage as a pedigree classification method and research tool both for evaluating aptitudinal trends in the Thoroughbred and for observing the relationship between pedigree and racing performance.  Academic intent just isn't the stuff that sells newspapers.

Now, almost a quarter century later, the press and the public continue to keep it SIMPLE.  As a result, whenever a horse does something on the track that appears to contradict conventional Dosage wisdom, it is regarded as a failure of Dosage and another nail in its coffin, so to speak.  This has been especially true in recent years of the Dual Qualifier (DQ) concept, originally introduced as a model defining common characteristics of American classic winners.  It, too, was made SIMPLE by the press, which focused on its literal definition and on the Derby almost exclusively because the other classic races lack a similar degree of public appeal.  At the same time, those students of Dosage who have learned how to use and interpret the information it provides continue to benefit.  Theirs is a victory of SUBTLETY over SIMPLICITY, of OBJECTIVITY over EMOTION.

It is widely known that after a remarkable span of almost 20 years in which DQs had dominated the Kentucky Derby, no DQ has won since 1997, although twice in that span (1998 and 2003) no DQ actually started.  The failure of any DQs to make the starting field had been unprecedented, at least since 1946.  In any case, there are no DQ Derby winners in the last six races in which at least one has started.  Of course no one bothers to mention that DQs failed to win the Derby in seven straight tries between 1965 and 1971.

Also ignored by the press in the throes Derby fever is the record of Dual Qualifiers in the other classics.  The following table compares the records of DQs in the Preakness and the Belmont in two separate timeframes - 1973 to 1989 and 1990 to 2005.  The data include the number of combined races, number and percentage of DQ wins, number and percentage of DQ starters, number of total starters and Impact Value (IV).  IV is a measure of success relative to opportunity assuming a random population.  In this case it is the ratio of the percentage of DQ wins to percentage of DQ starters.  If the percentage of DQ starters is 20%, and they win 20% of the races, the IV is 1.00.  This would indicate no advantage to a horse with DQ qualifications.  On the other hand, if 20% DQ starters win 40% of the races, the IV is 2.00, suggesting an advantage.  Complete DQ records are available by decade for each separate classic race here.

The Combined Record of DQs in the Preakness and the Belmont
(1973-1989 and 1990-2005)

The results are clear.  In contrast to DQ performance in the Derby, DQs have a superior record since 1990 as compared to their record between 1973 and 1989. 

It is certainly true that DQs in the Derby have done poorly in recent years.  The next table displays results for the Derby in the same format as for the other classics above.

The Record of DQs in the Kentucky Derby
(1973-1989 and 1990-2005)

The difference is dramatic, although the data since 1990 still show a positive IV.  However, the modest 53% advantage for DQs in the Derby in recent years pales in comparison to the huge advantage in earlier times.  This brings us to the critical question of why we see the observed decline.  The answer may lie in the evolution of aptitudinal type occurring in the Thoroughbred, an evolution that because of the truly unique conditions of the Derby expresses itself more fully in that race than any other.

We will address the problem in the following way.  Let's assign different numbers to represent the conditions of being a DQ and not being a DQ; say 0 and 1, respectively.  This is reasonable because DQ status is an on-off condition.  Then we will plot the zeros and ones against the year and generate a linear trend line to determine the best fit of the data.  Separately, we will use our existing plot of Derby winner DI vs year found here.  This chart shows, with its trend line, the undeniable increase in DI among Derby winners over the last 60 years and represents a steady and continual infusion of more speed into the pedigree of Derby winners over time.  We will then superimpose the two charts to see if there is a correlation.  In order to generate some separation of the trend lines, we can offset the DQ on-off numbers to 4 and 5 instead of 0 and 1.  The result of this analysis is shown in the following chart.

A Correlation of Derby Winner DI with DQ Performance

The DQ trend line covers the years 1973 to 2005 while the DI trend line covers the years 1940 to 2005.  The y-axis (0 to 10) is absolute for the DI data and relative for the DQ data.  Since we are only trying to show the relationship between the two over time, the degree of offset is not important.

What is remarkable is the virtually parallel relationship between the increase in speed in Derby winner pedigrees and the change in the nature of DQ performance.  One obvious conclusion is that the DI 4.00 guideline in defining a DQ may no longer have the relevance it once had.  The situation is made more complicated when we consider the results from our recent study using Beyer Speed Figures correlating increasingly inferior Derby performance with higher DIs of the winners.

So once again, and contrary to any media interpretation, Dosage is neither right nor wrong.  It simply reflects reality, allowing us to get a better handle on the evolution of Thoroughbred type and the consequences of that evolution.