29 November 2009
If you would like to comment on my letter and/or contribute your own ideas on how to fix the rankings please visit
and post a comment.
I also invite you to write to Johan Knaap at
and encourage him to fix the rankings.
And please forward this message on to your friends and colleagues who will be interested to learn about these issues.
During the last eighteen months several of my columns for an international equestrian magazine have called upon the World Breeding Federation for Sport Horses (WBFSH) to improve the methodologies used to create its rankings so the final product is useful to sport horse breeders. In these articles I offered specific suggestions on how to improve the rankings.
The WBFSH recently released the rankings for the 2009 competition year (see www.wbfsh.org). Unfortunately the suggestions I and other breeders offered you and the WBFSH have been ignored. Nothing has changed in the way the rankings are computed and we breeders are once again left with a ranking system that is of little value but are nevertheless the source of self-promotion among the studbooks that are ranked at the top.
Below I offer you a number of concerns and suggestions about the rankings:
1. The studbook rankings do not reflect the true genetic contributions of the studbooks.
Under the current system for computing rankings the studbook a horse is born into earns all the credit for the horse's success in sport. This means that the genetic contributions that studbooks make to each other are not recognized.
Please allow me to offer a few examples to illustrate my point. Studbook Zangersheide is ranked as the 8th best studbook for producing international showjumpers and is less than one-half of one percentage point behind the Oldenburg Studbook, which is ranked 7th. This is a great accomplishment for a relatively young studbook.
However, not a single one of the six competition horses used to compute Zangersheide's ranking was sired by a stallion born into Studbook Zangersheide. Instead these showjumpers were sired by Holsteiner (in fact three were sired by Holsteiners), Oldenburg, Belgium Warmblood, and KWPN stallions.
Further not a single one of the six competition horses used to compute Zangersheide's ranking is out of a mare born into Studbook Zangersheide. Instead these showjumpers are out of mares born into the Selle Français (two mares, on fact), Hanoverian, TB, KWPN, and Belgian Warmblood studbooks.
So in the case of Studbook Zangersheide its 8th place ranking is based on not a single sire or dam that was born into that studbook. From a genetic perspective, is Zangersheide really the 8th best studbook in the world? Or does its very liberal and open registration policy allow it to free-ride in the rankings on the genetics found in other studbooks?
This is not an isolated case. In your own KWPN Studbook, which was ranked as the best studbook for producing international showjumpers, three of the six horses were sired by non-KWPN stallions (Oldenburg, Selle Français, and Holsteiner). In the case of the 2nd place Selle Français Studbook, one of the six horses was sired by a Hanoverian stallion. The 3rd place Holsteiner Studbook had one international showjumper sired by a Selle Français stallion. The 9th place Swedish Warmblood Studbook has five of its top six showjumpers sired by non-Swedish stallions and the 10th place Danish Warmblood Studbook has all six of their top showjumpers sired by non-Danish stallions.
And in case you think this anomaly is restricted to the showjumping rankings, please note that much of the Irish Sport Horse Studbook's success in the eventing rankings has been built for the last decade on the genetics of a Holsteiner stallion, Cavalier Royale, which the studbook refused to approve until shortly before his death despite standing at stud in Ireland for about a decade.
And in the dressage rankings, your own KWPN Studbook, which is ranked first, has a similar predicament. Five of the six dressage horses whose results contributed to the KWPN's top ranking were sired by non-KWPN stallions: Gribaldi (sire of two horses in the rankings) and Partout are Trakehner stallions and Contango and Amsterdam are Oldenburg stallions. Shouldn't the Trakehner Studbook, for example, enjoy some of the credit in the WBFSH rankings for the amazing success in top sport by Totilas, Painted Black, and Nadine? Shouldn’t the Oldenburg studbook enjoy some of the credit for Ravel and Pop Art? I think they should.
The WBFSH must come up with a methodology that takes into account the genetic contributions made by stallions (and mares) that were not born into the particular studbook. If not, then the true contribution each studbook makes to producing world-class athletes will remain distorted and the winners will be those studbooks that are most adept are buying-in, rather than breeding, world-class stallions and mares.
2. The studbook rankings do not give credit to TBs.
The studbook rankings exclude thoroughbreds (presumably because TB sport horses are not born into a studbook that is a member of the WBFSH) and, therefore, under-estimate their true impact on sport horse breeding and sport.
The WBFSH must create a studbook ranking where TB sport horses can be classified and where the contribution of TB sires to other studbooks can be measured. For example, although the TB showjumping stallion Favoritas xx appears in the 2009 ranking as the sire of a showjumper, a TB international showjumper like Favoritas xx could not appear in the studbook ranking because no classification exists for a TB studbook. And the great TB stallion Heraldik xx, which sired horses in all three Olympic disciplines (a truly amazing feat), should have his contributions to other studbooks credited to the TB studbook, as recommended above in point 1.
3. The studbook rankings penalize small studbooks.
The studbook rankings are computed by summing the points earned by the top six competition horses (in each discipline) born into each member studbook. The fact that some of the studbooks produce 10,000 – 15,000 foals each year while others produce much smaller numbers is not taken into account.
I have already pointed out the immense, but unrecognized, contribution the Trakehner Studbook has made to the KWPN's first place ranking in dressage. In 2009 the Trakener Studbook registered 1,234 foals. Let's assume the KWPN registered 12,000 foals – approximately 10 times as many foals. Now let's take the KWPN's 12,200 ranking points and divide that by its 12,000 foals: that gives us 1.02 points per foal. Now let's do the same for the Trakehner Studbook: 5,498 ranking points divided by 1,230 foals gives us 4.47 points per foal. So the Trakehner Studbook is producing 4.5 times as many points per foal as the KWPN Studbook. Which is the better studbook for dressage? Now the answer is not so clear.
Of course this is a rough and ready calculation but it does illustrate that we cannot understand the true genetic value of a studbook if the WBFSH ranking methodology does not adjust for the size of the population of each studbook.
The WBFSH must come up with a methodology that takes into account the size of the studbook and the number of chances it has to produce top athletes. If not, the rankings will remain biased in favor of big studbooks and biased against small studbooks. One solution could be to count the international results from every competition horse born into each studbook and adjust for studbook size by computing two rankings based on the average number of points earned (a) per international athlete and (b) per horse born into the studbook at some point in time (for example, ten years prior to the ranking year).
Until now in this letter I have focused on studbook rankings. Now I would like to turn our attention to the ranking of sires.
4. The sire rankings penalize younger sires.
The WBFSH ranking of competition horses only uses data from upper-level international competition so the rankings are skewed toward established sires whose progeny are old enough to compete at those levels. So for all practical purposes the highly-ranked sires will be 15 years or older and, in fact, many of the sires at the very top of the ranking will be aged, dead or have age-related fertility problems.
The WBFSH must change its methodology to include results from all international competitions so we can identify up-and-coming sires of international athletes. Further, two age bands should be created and sire rankings should be generated for each band: sires of international athletes 6 - 8 years of age and sires of international athletes age 9 and above.
5. The sire rankings over-reward popular and fashionable stallions.
The sire rankings are computed by summing all the points earned by all the progeny of each stallion. Stallions like Darco (ranked 1st as a sire), which sired a very large number of progeny, have a built-in advantage over sires like Mr. Blue (ranked 11th), which sired relatively few foals.
The ranking of sires must take into account the number of progeny produced by each stallion. A sire that has had 2,000 foals has had many more chances to produce international athletes than a stallion with 500 foals. Ideally this new ranking would take the total points earned by each sire (the basis for the conventional ranking produced now) and divide it by the number of progeny that are over six years of age. (I recommend six years of age because that is the youngest that we would normally see a horse competing in international sport.)
Along with this change to the rankings the WBFSH should also report the number of progeny age six and over; the average number of points earned by each progeny age six and older; and the strike rate, meaning the percentage of progeny age six and older that have become international competitors.
6. The sire rankings have no memory.
In years past, when the WBFSH rankings were published in book form, two rankings were computed: a current year ranking of sires and a ranking over the past ten years. As the WBFSH migrated from book to DVD to now online publication of the rankings we lost the longitudinal data and rankings. This is very unfortunate.
The WBFSH must publish rankings over a longer time period, such as the aforementioned ten-year period, so breeders can see the contributions of sires over time. These rankings should have the same corrections and adjustments as noted above.
7. Dam-sires are ignored.
It would take a computer programmer or statistical analyst less than one minute to write the code to create a ranking of dam-sires. The ranking should be based on the same methodologies noted above.
Finally, let us turn our attention to the ranking of breeders.
8. Breeders are ignored.
When the WBFSH rankings were published in book form the name of every sire and competition horse was accompanied by the name of its breeder. For the last few years we have had a bizarre state of affairs in the World "Breeding" Federation for Sport Horses: the breeder is no longer recognized in the rankings.
The WBFSH must print the name of the breeder of every sire and competition horse and once again rank breeders. It would be very helpful if the breeder ranking were for the current year and ten-year periods.
Johan, I now invite you to consider these points and to reply. I believe you have a fiduciary responsibility to the WBFSH and, by extension, to the members of every studbook that is a member of the WBFSH, to improve these rankings so they are useful tools for breeders. Until now that has not happened.
As I see it, you have three choices: (1) Fix the problems; or (2) Debate me and others on the problems and the solutions; or (3) Resign from your position of authority over these rankings. The one thing you cannot continue to do is ignore the problems.
I will gladly debate my points with you in any forum including the blog I have established specifically for this purpose:
Thomas Reed, Ph.D.