Guest Columnist of the week is Dr. Maren Engelhardt. She was born into an equestrian family and has been involved with horses as a rider and breeder for all her life. Together with friends, she created www.trakehners-international.com and writes for international magazines such as Der Trakehner, Horse International, The American Trakehner, Warmblood Magazine and others. She lives in Germany.
Numbers Don't Lie, or Do They?
We modern breeders do love statistics: there is no question about it. We have a plethora of index charts, sport results, WBFSH rankings, breed association publications, the mighty BLUP and more to chart the next outstanding match between our mare and the “best” stallion we can find. Most of us have no clue which mathematical models were used to reach conclusions on numbers – most of us never notice where the methods are flawed. In fact, some of us actually don’t care about numbers at all. For all the (pseudo-) scientific babble out there, breeding horses is still very much a “gut” thing and the success of several centuries of selective horse breeding really can’t be all that wrong.
The Stallion Performance Test
One part of the stallion selection process, and a very important one, is the stallion performance test. The idea was born at Trakehnen and also at the State Stud in Celle, pretty much around the same time in 1926. The idea is simple: test the young colts before you allow them to propagate. At the East Prussian test station in Zwion, young stallions were ridden, fox hunted, driven, and assessed six ways from Sunday for a full year before a select few had a shot at breeding. Obviously, back then, people had the time. But now things have changed, and the luxury of a long test is now passé. For many years, the German 100-day test was the epitome of stallion testing – copied worldwide. Those 100 days got cut to 30 + 70 a while ago, mostly for animal welfare reasons. The jury is still out on that.
Traditionally, all issues relating to breeding animal selection and testing were a State-run process in Germany. That changed only recently and led to the German Equestrian Federation (FN) and Germany’s 25 organized breed registries taking over control of the process. Now the FN has come up with a new index number for all of us to ponder. But let’s take a look at the history and meaning of the conventional index first.
Most of you are aware of the German “index system”, where stallions are ranked based on their “value” as a breeding stallion. Breeding indices are estimates of the hereditary performance potential of a horse, which it will pass on to his offspring. The goal of that index is to differentiate these hereditary performance traits as accurately as possible, so that breeders may have a reliable basis for breeding decisions. In other words, the index is supposed to give you an idea of how likely a sire is to produce a successful competition horse.
In the old days, this system relied heavily on offspring competition data – success at S-level usually translated into a higher index of the sire of those performing horses. Because of the way it was calculated, a larger number of performing offspring produced a higher index for an individual stallion. The system was split into dressage and show jumping. And that is where the first major misstep was taken. One may ask what happened to eventing? Those horses don’t compete? Is there nothing to learn from data based on truly versatile athletes? No chance the other two big disciplines might actually profit from including such data? The arguments against an eventing index range from “low numbers of shows and hence limited data sets”, “there is no money to create a new database (!)” or “the discipline is “murky” – not really dressage, not really show jumping”. Well, that’s kind of the point … which begs the question.
So consequently, an astonishing number of highly successful, reproducibly reliable sport horse sires (often TBs, Anglo Arabians, “special blood” stallions including the Trakehner) are not even noted. The leading WBFSH sires in eventing are shockingly absent from the German index system. Interestingly, they do get their FN dressage and jumping indices. Well guess what those numbers look like.
Here is a good example: in 2010, stallion X had roughly 260 competing offspring in the German system. This stallion’s offspring lifetime earnings are nearing 800,000 euro – and he didn’t stand at stud all that long. He has over 15 approved sons, over 350 registered broodmares in various German registries. Our mystery stallion has produced international Grand Prix dressage horses, top quality show jumpers and a staggering number of international championship medal-winning event horses. Yet his FN dressage index is 93, his jumping index is 99. Would you consider such a stallion if your tool to pick was the coveted FN breeding index? I would think, most likely not. What a big loss for the sport horse world in the long run! The stallion? Heraldik xx.
Here is another example: imagine you are looking at numbers only, not names, and are given a choice of sources. You see a stallion that is a coming 9-year old. His own lifetime earnings amount to 900 euro (stallion never made it past lower levels in dressage). This is a highly popular stallion with a full book each year, yet in 2010, only 5 of his 11 FN registered offspring were competing. Of course it would be unfair to expect them to compare to offspring sired by older stallions, but even so, their combined net income of 423 Euro in 2010 is meager at best. What a contrast that is to the fact that this same stallion has produced 25 (!) approved sons so far. But guess what? – the name sells. This is Quarterback.
I wish somebody would create a system to compare stallions in a more fair way, based on the number of foals that were produced in total, compared to how many of them ended up in competition. Then include in the calculation all “real” competition results, from the lowest levels to the advanced levels. And even that is a number which will miss a major point in horse breeding: the fact that just the sum of genes in a particular animal means little after all. It’s not so much about which genes you have – it’s more about environmental factors that “make” a horse. There are stallions out there that have had very limited books, for a variety of practical reasons, such as their own extensive performance careers. And yet a significant proportion of their offspring made it to the upper levels in competition. Those are the ones we should aim to identify. And that will not happen with any of the current index systems used worldwide.
Changes for the Better?
The other major damage to Germany’s index system came with a set of changes established several years ago, when true offspring performance data was devalued, and other parameters were added to the formula: success of daughters in mare performance tests (done at age 3 and 4 usually, so a far cry from “real” performance), success of sons at stallion performance tests (ditto), the Bundeschampionate rankings, and the indices of at least 2 generations of ancestors. Consequently and literally overnight, some of Germany’s most proven, reliable and successful all-time great sport horse sires (e.g. Argentinus) disappeared from the top of these index lists and were replaced by younger stallions. Unproven as sport horse sires at best, but stellar young sires with a number of young horse show successes under their belts. Nobody says that the latter are bad horses. What needs to be pointed out is that they are unproven due to age, and their “value” to the betterment of sport horse breeding is still a big question mark. Yet modern stallion promotion would have you believe that the great index system points the way. It is a fact that the Top 50 stallions in any given FN index include some names that certainly belong there, but it is also a fact that some true losers end up in this list as well. How does one know which “statistic” is valid?
And if you think this is not a big enough mess already, take into account the most recent “change” to make things “better”. This amendment seems like one with the quality of a kiss-of-death for a fair number of stallions and, unfortunately, it will be mostly those of “special blood” – a group we sorely need!
The new stallion performance test reform in Germany now has the advantage that scores are comparable between stations, something that didn’t exist before. So how does this work?
A stallion is given scores throughout the time he spends at a test station, using the traditional criteria such as temperament, constitution and willingness to perform along with conformational qualities (gaits), jumping, and in the 70-day format, some cross country work. But instead of going home with scores and points, the young stallion now goes home with a combined index for his “breeding value”. This index is comprised of the scores from the actual test and also of his “ancestor breed value”. The final breed index is the one and only number that will determine the stallion’s future career, thus muting objective measurement of that stallion. The 25 breeding registries in Germany that were all part of this new reform have set their minimum standards for full approval quite disparately, ranging all the way from 80 to 120.
Is it a surprise that a very vocal discussion about this grand new system has sprung to life in Germany? What is being overlooked is that the system is now built on a flawed foundation. How is that going to help us produce better sport horses? Logic dictates to get rid of the Heraldiks in our pedigrees and use the Quarterbacks. For the betterment of sport horse breeding? We are creating a virtual list of born breeding index annihilators, even before anybody ever puts a saddle on their backs?
Here is how this new system is particularly disastrous for blood horses:
1) Many of these stallions breed very small numbers of mares each year. They are the choice of breeders that maybe hope for a filly to keep as a future brood mare, or by breeders specializing in event horses. Let’s assume that Lauries Crusador xx is the famous exception from the rule. Now, a stallion with even a high number of money winning offspring in eventing gets zero recognition for his achievements. There is no index for this discipline.
2) Ancestors of blood horses or full TBs are usually not ranked in any of the classical riding disciplines. Their success comes from the racetrack. Hence, they can’t help the ancestor index, again, this part of a stallion’s history remains blank. I’m not suggesting to start counting racing success per se, but blood horses contribute traits that we all need: soundness, stamina, character, ideally documented by a long career on the track. None of these important factors find a way into the index.
3) Many of these special stallions are “alien residents”. They were born outside Germany, have ancestors that never set foot into Germany, and possibly also have (successful) offspring that are not in Germany. Hence, they don’t count towards any type of index.
Three strikes that put a number of stallions into a huge disadvantage even before they breed a single mare. Now, how likely is it that stallion stations are going to invest into blood horses now that we have a system in place that makes life even harder for them?
Well-known journalist Claus Schridde put it in very clear words: “Surgery a success – patient dead” was the title of a recent article that appeared in the German magazine Züchter Forum. “If despite excellent under saddle results, a horse cannot achieve a positive breeding value index, the new system is a joke,” he wrote. Thomas Casper, one of the most successful stallion managers in Germany and owner of Gestüt Birkhof (former home of Heraldik xx), said that “the true test scores need to be the emphasis, not an ancestor index that is highly unreliable, and especially biased against the blood horse.”
In the real world, this has tremendous impact on stallion sales, auctions, and those breed registries that do not produce enough for the big numbers. The Hanoverian Verband for example has set the mark for Hanoverian stallions at 100, for all others at 110 to be granted full approval. Those numbers are virtually impossible to reach for most blood horses and a large number of smaller-registry stallions. Stallion owners are now afraid that without a chance of approval in Germany’s biggest book, their breeding stock will have even less business – how likely is it now that big stallion stations will buy into smaller breeds - those that without a doubt had profound impact on sport horse breeding (think French Anglo Arabian in Oldenburg or Holstein, or Trakehner in Hanover) - yet without a chance to recoup the investment? Word has passed around that certain big stations are already working on a “breeding value list” for the 2012 crop, meaning only colts where the ancestor index is high enough will be bought for future stallion careers. We’re talking about weanlings here. Nobody has even a remote idea how well they will be performing under saddle in 2015, and already we’re selecting, based on a more than shaky index system.
Straightening Out the System
Apparently there was already enough loud crying, especially from the breeders of upper level horses and eventers. Just founded the new system is already in “fixing” mode. Yet instead of getting rid of an index that’s just not telling us much, blood horses are now to receive “special consideration” - of the kind where hand-picked judges meet at the test stations to check out those stallions individually. So instead of creating a system with less room for human manipulation and the risk of harmful corruption, we now set the stage for more.
Heated discussions between the representatives of FN, breed registries and the German Stallion Owners Association are underway. And what is the direction now? It is unlikely that the breeding value index will be eliminated right away. In the end, the registries have the final word on how this should work out. So us breeders, right? We’re the paying members. Yet somehow I feel utterly misrepresented. I may be one of the odd idealists that is into breeding blood horses for eventing (God forbid), but I’m far from alone. And it doesn’t stop with TBs. It will impact pretty much every stallion not sired by the Big Five plus 1 (W,D,R,S,F+Q) and it will have profound impact on the already alarming rate of gene pool restrictions we see today. Open the catalogues of the most recent Oldenburg, Trakehner or Westfalian approvals and witness it first hand. I’m not even mentioning Holsteiners. This is a very slippery slope we’re on, all of us, no matter if in Europe or elsewhere. In the light of mind-boggling auction prices, more bizarre movement extravaganza with each stallion crop, our own tendency to “ooh” and “ahh” about the latest flavour of the day, we should really sit back and think this one through. We need stallions that as individuals have proven to be successful, in competition and/or as sires of successful competition horses. What we don’t need are stallions with their “merits” based on an elaborate mathematical model, promoted as the next best thing after the invention of the wheel.
“Ultimately, many of the German and FEI iterations of "value" ratings have erred in overstating what one can conclude from the numbers. It is just about trying to make these ratings into something more than they really are. Defusing that error would go a long way in helping people use the numbers correctly,” says Timothy Holekamp, my good friend. From the perspective of this (very successful) American sport horse breeder, it looks as if “all of this ranking, rating, valueing, and what-not is classic German culture ... an attempt to reduce art into science, or perhaps to deny the ethereal aspects of a practical process that is mistakenly described as "art" when it is actually a gnostic body of knowledge that no one is able to articulate/teach/record.” Now just to find a way to combine the two.
Happy breeding season 2012, everybody!
- by Dr. Maren Engelhardt for Eurodressage
(Dr. Engelhardt would like to thank Melitta Burger and Dr. Timothy Holekamp for valuable discussion of the topic)