[This is the fifth and final part of the second lecture on the Rasmussen Factor. Comments and observations from various readers have been appreciated.]
Quote in Full Context, Implications, Conclusions
I began this lecture with an unidentified quote. That quote was from Racehorse Breeding Theories, Chapter 16, written by Ross Staaden B.V.Sc., Ph.D. I will now amplify on that quote and give it a more complete context. Here is the complete text from Staaden:
“Also I will quote Dewey G. Steele. The man was regarded as one of
Kentucky’s most eminent authorities in the field of horse genetics.
Nelson Nye, the Quarter Horse writer, quotes him in ‘Speed and the
Quarter Horse’ (1973, page 49). Steele compared two extreme groups,
the stakes winners of 1935, 1940 and 1941 with last horses in the
last races on Thursday and Friday afternoons of the same years. He
used some statistical analysis. Inbreeding in the two groups was
compared, including the relationship to various ancestors, St. Simon,
Bend Or, Domino, etc.
” ‘. . . There did not appear to be any significant differences
between the two groups. It was apparent that the poors trace rather
quickly to the same ancestry as do the best performers. The greatest
ancestral difference between these two extreme groups was in the
first generation. It declined rapidly in each of the succeeding
generations. . . .”
” ‘Evidence from these studies indicates that pedigrees should be
judged primarily upon the basis of the first and second generations
and that ancestors beyond the third generation may for all practical
purposes be ignored.’
“The famous racing writer Joe Estes came to a similar conclusion, due
to his research published in The Blood-Horse. ”
About this time three years ago I had occasion to discourse with Preston Madden. Mr. Madden told me that he took a class from Professor Dewey G. Steele at UK in the late 1950s and that it was the only class he never cut while he was at UK.
These are hard acts to follow: Steele, Uncle Joe Estes, Madden. Yet I suppose I should add my own two cents’ worth as to the implications of these results on the RF factor. The reason that the RF factor did not fare any better than it did seems pretty obvious to me: most of the duplications involved are back in the fourth and fifth generations.
Just because Northern Dancer (or Mr. Prospector or Secretariat or Halo or Almahmoud or Somethingroyal or whoever) is great in the first generation does NOT mean that when this name appears in the fourth or fifth generation, it carries with it a positive “influence.” This is the fatal flaw in RF, all other inbreeding schemes, and all pedhead thinking in the first place. Pedheads see a great name in the fourth or fifth (or 20th) generation and ASSUME that it carries with it a positive “influence.” Based on ZERO proof whatsoever.
As you go back through the generations of a pedigree, the names in the pedigree generally get better and better (which is sort of what Steele was saying). Ergo, the farther back in a pedigree you go, the more the great names are all competing with each other, and for every winner there is going to be a loser. Basing contemporary expectations or pedigree evaluations on fifth-generation evidence is a zero sum game.
Perhaps another baseball analogy would be useful here. Baseball players are not expected to perform (much less excell at) all nine positions. A sire may appear in eight possible positions in the fourth generation. Most sires do appear in all eight positions (which would be like a baseball player trying to play all nine positions; inevitably he would be good at some and not so good at others). There are a few exceptions to this rule. Prince Rose, for example, had very, very few daughters surviving into modern pedigrees. Ditto for Discovery, who had very, very few sons surviving into modern pedigrees.
Pedheads ASSUME (based on ZERO proof) that great sires will have a positive “influence” at each of those eight possible positions. The reality, as determined by statistical means, is that very, very, very, very few sires, no matter how great they are in the first generation, are good at all eight positions in the fourth generation. They will be good at some positions and bad at others. Their overall “influence” may be positive or negative (the latter a lot more frequently than you might think, and the reason is that they are competing against other equally great or better sires).
Take a look at Chapter 9 of Racehorse Breeding Theories. Look specifically at Table 2, which shows the 20 most popular sires in the fourth generation of North American-bred foals of 1995-97. The norm for that group was 3.74. Can you find any of those 20 sires who was good (had a positive “influence,” was above 3.74), at all eight positions? No, you can not. And that is par for the course.
What does this have to do with the RF factor, which is about “superior” females, not males. Well, the vast majority of the duplications among these RF qualifiers involve sires of some sort: Almahmoud being the second dam of Northern Dancer and Halo, Somethingroyal being the dam of Secretariat and Sir Gaylord (and others), Flower Bowl being the dam of Graustark and His Majesty, etc. If the sires are no good at these points of duplication, obviously the duplications are not going to do anything “positive” for the pedigree.
Duplicating names in pedigrees (male or female) does not make any sense unless you know for a fact that the duplications involved are “positive” in the first place. Which no one knows unless they have looked at a couple hundred thousand pedigrees and done the homework on it diligently and dispassionately and with an open mind (which very few people evidently possess, especially among those who profess to be students of pedigrees).
It might be interesting to take this same group of 54,244 sales foals and examine them for inbreeding to Northern Dancer, for example. It would be interesting to see how that group stacked up against the RF qualifiers, whether that group’s results justified its prices (which RF did not). Perhaps, that is an idea for another day.