Observations on Thoroughbred Evolution

The adage "nothing endures but change", a quotation attributed to the Greek philosopher Heraclitus (c.535-475 BC), is as true in Thoroughbred racing as it is in any facet of life.  Most of us who have followed the sport long enough to have developed a sense of its history may easily perceive that things are not as they once were, but many may have a harder time appreciating the nature of the changes they have seen.  For example, we know that medication, both legal and illegal, has become more sophisticated and probably plays a greater role than in the past.  Yet how well do we understand the depths of its effect on the horses and on the races in which they compete?  We have seen legendary races disappear from the scene in favor of new events, or their distances shortened.  Those in charge argue such changes are needed and are made in racing's best interests, often to the dismay of purists, traditionalists and nostalgia buffs.  And we have observed once dominant sire lines fade into obscurity to be replaced by other emerging bloodlines, either providing breeders with fresh opportunities or, depending on your point of view, demonstrating an erosion of the breed's vigor.  Regardless of our feelings about them, the changes occurring in Thoroughbred racing are ongoing and constant.  Change is the human condition and it is in our interest to understand the forces of change so that we may establish continuity between the past, the present and the future.

Thirty to forty years ago Thoroughbred racing was dominated by a dark bay stallion named Bold Ruler, an elite, classic-winning racehorse and an even better sire.  In the 1960s he led the general sire list an unprecedented seven straight times and then once again in the 1970s.  No other sire in the 20th century could match that feat.  He was the leading freshman sire of 1962 and he topped the juvenile sire list six times.  Along the way, Bold Ruler got an amazing 23% stakes winners including eleven champions, among them perhaps the greatest of all, Secretariat.  To claim that the roster of his highly successful sons at stud is impressive is an understatement: Bold Bidder, Bold Commander, Bold Hour, Bold Lad, Boldnesian, Chieftain, Cornish Prince, Dewan, Irish Castle, Jacinto, Raja Baba, Reviewer, Secretariat, Top Command, What a Pleasure and more.  This group sired classic winners and Grade/Group 1 winners in America and Europe and many individually were prominent on or led various year-end sire lists.  Bold Ruler's is a compelling story about a very special Thoroughbred.  However, within just two or three decades from the time Bold Ruler reigned supreme, we have arrived at a point where only one of his living direct male line descendants legitimately can be considered a world-class sire, A.P. Indy by Seattle Slew by Bold Reasoning by Boldnesian by Bold Ruler.  Not only do things change in the racing world, they can change in what seems like an instant.  No one can know for sure what will be the fate of more recent dominant sires such as Northern Dancer and Mr. Prospector.  Whether their influence contracts as has Bold Ruler's, or whether they expand their influence and ascend the heights to become the new Phalaris is something we will learn over time.

A topic of active discussion during the time Bold Ruler, his sons and his grandsons were churning out one high class runner after another was how well they were crossing with mares from various broodmare sire lines.  Bold Ruler/Princequillo, or more broadly Nasrullah/Prince Rose, was considered a successful "nick" (or bloodline cross), exemplified by superior performers such as Secretariat, Bold Lad, Successor, Revidere, Seattle Slew and a host of others.  Whether that cross was something special is a matter of conjecture.  Nevertheless, the interaction of bloodlines was and still is of broad interest to owners, breeders and fans, making it an area of study worth pursuing.

The Bold Ruler/Princequillo and Nasrullah/Prince Rose designations are part of an even broader bloodline cross designation: Phalaris/non-Phalaris.  "Phalaris" labels every Thoroughbred that is a direct male line descendant of that early 20th century superstar sire while "non-Phalaris" labels everything else.  To put Phalaris in perspective, over 90% of North American graded stakes winners born in the new millennium are his direct male line descendants.  In the history of the modern Thoroughbred there was Eclipse and then, over a century later, there was Phalaris.  No other sires have come close in terms of long-term influence on the breed and no one can accurately predict which later sire will be the next in line.

Breeding, of course, is not limited to Phalaris line sires mated with non-Phalaris line mares (for convenience, we will now refer to the Phalaris line as P and the non-Phalaris line as NP).  All combinations exist in the Thoroughbred: P sires with P mares (designated P/P), P sires with NP mares (P/NP), NP sires with P mares (NP/P) and NP sires with NP mares (NP/NP).  As we might expect because of the innate differences among the individual sires comprising the separate bloodlines, each combination represents a unique set of traits and characteristics, and each combination assumes a competitive position against the others in the struggle for supremacy.  Since the selection process of Thoroughbred evolution is driven primarily by the requirement of success on the track, combinations of bloodlines having traits and characteristics best meeting that requirement will prevail.  The specific demands ultimately depend on how racing is configured at any point in time, and the selection process for bloodline crosses should be no different than the selection process for individual sires.  The ones best suited to compete in the current racing environment will survive while those less suited will eventually disappear.

It is critical to recognize, however, that as bloodlines or bloodline crosses evolve, their traits and characteristics also evolve.  The direct sons and daughters of Phalaris, born almost a century ago, may bear little resemblance in type, structure or disposition to direct male line descendants of Phalaris born eight or ten generations later.  As a result, when discussing P/P, P/NP, NP/P or NP/NP it must be in the context of time because the very meaning of evolution is change over time.  P/P in 1980 isn't necessarily the same thing as P/P in 2006.

Some people attribute various trends in Thoroughbred performance to the appearance of bloodlines or their combinations in a pedigree.  The proposed association almost always ignores the dynamics of the evolutionary process and treats the bloodline or bloodline cross as if it is a static thing, never changing qualitatively, and independent of time or place.  Unfortunately, the unique and differentiating characteristics of a bloodline or bloodline combination are rarely if ever measured, meaning that the evolution of those characteristics proceeds unobserved.  Therefore, at a given time, any number of factors could be responsible for the noted trend since we have no idea what is genuinely responsible for it or, truthfully, if it is even real.  As a consequence, invoking P/NP, for example, as a singular explanation for an observed historical trend is specious and recalls Shakespeare's words in Macbeth of "a tale told by an idiot, full of sound and fury, signifying nothing".  

One easily measurable trait, however, and one that has actual physical significance is the average winning distance in furlongs (AWD) associated with a sire, a bloodline or a combination of bloodlines.  AWD is a direct reflection of the kinds of races won by horses that have been shaped by specific pedigree influences. 

For our purpose, then, we can examine the AWD of the four possible P and NP crosses over time in an attempt to observe variations that may have occurred and which may indicate that P and NP in 2006 don't have the same meaning as P and NP in 1990 or 1980 or before.  If we are able to confirm significant changes, then conclusions drawn about the impact of a static P/P, P/NP, etc over a span of years are invalid and misleading.

Consider the following quote from the web site http://www.sportsnetwork.com/default.asp?c=sfgate&page=horse/kentucky/2005/amer_turf.htm.

"In the last 36 years (1966-2001), 22 Kentucky Derby winners had the "Phalaris sire line/non-Phalaris broodmare sire line" pattern. This pattern made up approximately 35 percent of the starters and 61 percent of the winners, which translates to a solid impact value of 1.74."

In the cited study, "Phalaris" and "non-Phalaris" simply are names applied to sire lines, nothing more.  They are names that evoke no understanding of what they represent in a physical or biological sense.  Nowhere are the traits or characteristics of P or NP addressed, so unless they were the same traits and characteristics in 1966 as they were in 2001, the results are essentially meaningless.  Let's use an illustration to make the point more clearly.  The illustration is an exaggeration to be sure, but it is representative.

Carson City and Nijinsky II are both from the P line.  Meadowlake and Le Fabuleux are both from the NP line.  For the sake of argument, let's say that 25 years ago the P/NP cross was typically embodied by the Nijinsky II sire/Le Fabuleux broodmare sire pattern.  Also, let's say that today the P/NP cross is typically characterized by the Carson City sire/Meadowlake broodmare sire pattern.  This is how the respective crosses look in terms of aptitudinal type as captured by AWD where the data are based on the pedigrees of winners of North American open stakes races since 1983:

Nijinsky II (AWD 9.39)/Le Fabuleux (AWD 9.16)

Carson City (AWD 6.62)/Meadowlake (AWD 6.99)

If most P/NP types 25 years ago were like Nijinsky II/Le Fabuleux, the probability of a P/NP classic winner would be reasonable considering the suitability for classic distances.  In stark contrast, if most contemporary P/NP types were like Carson City/Meadowlake, the probability of a P/NP classic winner would be remote at best.

Although extreme, it should be obvious to anyone that the disparity highlighted in the illustration between what P/NP may have meant 25 years ago and what P/NP may mean today is huge.  The truth is that the divergence is not nearly as striking.  Nevertheless, as we shall see, the actual difference is real and substantial and confirms the proposition that developing a statistical conclusion based on something like P/NP that spans a period 35 years is ludicrous.

The following table summarizes AWD data for all four possible P and NP crosses for each year since 1983.  The data are based on North American open stakes races (graded and ungraded) that were won by graded stakes winners, and the table includes, in addition to the AWD information, the total number of stakes races won each year by the graded stakes winners from each combination.  The emphasis on graded stakes winners is predicated on the certainty that the direction of Thoroughbred breeding, and consequently the breed's evolution, is determined by the highest class of racing stock. 

The table indicates that between 1983 and 1987 there were 803 open stakes races won by P/P horses with an AWD of 8.16 furlongs.  Over the same timeframe there were 1068 races won by P/NP horses with an AWD of 8.58 furlongs, 517 races won by NP/P horses with an AWD of 8.51 furlongs and 557 races won by NP/NP horses with an AWD of 8.75 furlongs.  P/NP was easily the dominant bloodline cross of that era.  By far the shortest AWD was reserved for the P/P horses while the longest AWD was associated with the NP/NP horses.  From this data alone it is reasonable to suggest that in the mid-1980s P was typically a speed line and NP was a stamina line.  If true, then combinations of P and NP, i.e. P/NP and NP/P, would be expected to fall between the extremes, which they do.

Now let's move ahead to another five-year period almost two decades later, 2001 to 2005.  Here we find spectacular growth in the number of races won by P/P horses to an astounding 2110 as well as an increase in AWD to 8.30 furlongs.  There have been significant changes among the other combinations as well.  P/NP has fallen dramatically to 674 races won and its AWD has undergone an equally precipitous drop to 8.19 furlongs.  NP/P has experienced a similar transformation, falling to 234 races won with a parallel decline in AWD to 8.21 furlongs.  Nominally, the AWD for NP/NP has actually increased to 9.01 furlongs, although the sample size makes the result questionable as the number of races won is a mere 60.  In fact, in 2006 through the month of September there had been only one North American graded stakes winner bred on the NP/NP pattern - Bishop Court Hill, winner of the Grade 1 Carter Handicap and by Holy Bull out of a Cormorant mare.  

In less than twenty years P/P has easily displaced P/NP as the dominant bloodline cross among North American graded stakes winners.  The displacement has advanced to such an extent that it effectively precludes a major P/NP resurgence.  What seems most likely is that not only P/NP but all bloodline combinations involving NP will continue to decline in the face of evolutionary pressure from P/P.  Just as the Bold Ruler sire line has lost its preeminence, P/NP is slowly but inexorably fading from the scene. 

Note that the total number of races within each five-year period differs by only about 4 1/2%.  Therefore, the number of races won by runners from the individual bloodline crosses closely mirrors the percentage within the entire population.  Between the mid-1980s and early 2000s, P/P winners demonstrate an increase from just 27% of the population to an astounding 69%.  The other three combinations, each involving at least one NP line, have all declined considerably - P/NP from 36% to 22%, NP/P from 18% to 8% and NP/NP from 19% to only 2%.  At the same time, the AWD for all races, independent of the bloodline cross, has declined from 8.48 to 8.28 furlongs, a phenomenon that is worrisome to many who criticize the direction Thoroughbred racing appears to be headed.

With no other information we can see that within a short twenty-year span, the type and influence of the four P and NP combinations have undergone highly visible evolutionary, if not revolutionary changes.  The impact of these changes is even more vivid when we visualize what has happened.  We can do this by charting the trend lines associated with each bloodline cross.  The data from the table have been graphed and the separate data series (i.e., P/P, P/NP, etc) have been subjected to linear regression analysis so that we may plot the trend line associated with each one.  Chart 1 displays the trend lines for AWD while Chart 2 displays the trend lines for the number of stakes races won.

Only the P/P trend line (blue) in Chart 1 indicates an evolution toward stamina while the three other P and NP combinations drift in the opposite direction toward speed.  The trend line for the breed overall is also toward speed.  Note the crossover point of P/P and P/NP occurred between 1999 and 2000.  And although the NP/NP trend line (green) still shows the most stamina, it should be understood that the paucity of data available for that cross in recent years makes an interpretation of the data more difficult.  In any case, for all intents and purposes, NP/NP is, in a word, over.  This assertion is easily confirmed by the trend lines found in Chart 2.

In Chart 2 we see patterns for the number of stakes races won.  These patterns are similar to those found for AWD in Chart 1.  Here the momentum favoring P/P is even more pronounced and the trend line for NP/NP has crossed the zero line.  And not only have the various P and NP combinations been rearranged in terms of relative influence on the breed, they are clearly evolving into transformed aptitudinal types.

When we analyze the affinity for turf racing expressed by the four P and NP crosses since 1983 we see comparable trends.  In this case, the overall percentage of stakes races on turf has increased appreciably and essentially all of the increase is due to a major shift upward in P/P coupled with modest shifts downward for the other three.  The trends are captured visually in Chart 3.

From the chart we can see that in the mid-1980s the best turf cross was P/NP and the worst was P/P.  In fact, all three crosses involving an NP line were superior to P/P.  Once again we have witnessed a distinct reversal of type and P/P has become the superior turf bloodline cross.

The last relationship we will examine is that between bloodline crosses and early maturity.  For many years it has been the conventional wisdom that the speed associated with the Phalaris line played a major role in the success of Phalaris-line stallions as sires of precocious two-year-olds.  As little as twenty years ago that appears to have been true, as suggested in Chart 4 which shows the trend in percentage of stakes wins by juveniles for each P and NP combination.  Here, too, the situation is changing, with P/P no longer noticeably dominant and also exhibiting the greatest rate of decline.  This is consistent with the enhanced stamina of P/P-bred runners thereby limiting opportunity in many of the shorter two-year-old races.  At this stage, there is little to distinguish among the different combinations with regard to early maturity.

On the basis of the data presented here it should be abundantly clear that when describing the characteristics of a bloodline cross, or "nick" if you will, the conclusions you draw and the significance of those conclusions have meaning only within extremely narrow timeframes.  The unique evolutionary pressure applied to Thoroughbred racehorses apparently engenders a rapid rate of change within the breed, making any assertions about long-term bloodline influences superfluous.