“Influence” (One Pedigree at a Time)

“Influence” is a tricky word. It means different things to different people. It means one thing to geneticists. It means another (entirely different) thing to all the rest of us. In a later post I will explain what “influence” means to geneticists. In today’s post I will try to explain what “influence” means to all the rest of us.

“Uncle Joe” Estes invented the average earnings index (AEI) back in 1948. Applied to sires, it remains a very viable statistical tool today.

AEI as a measure of sire performance is a measure of sire “influence.” An AEI of 1.00 is average. An AEI below 1.00 is below average and means that the sire possessing it has below-average “influence” (or is a negative “influence,” if you prefer). An AEI above 1.00 is above average and means that the sire possessing it has above-average “influence” (or is a positive “influence,” if you prefer). For example, A.P. Indy has an AEI of 2.90. That means his “influence” as a sire is 190% above average.

AEI is based on average earnings per starter per racing year. A similar index called the standard starts index (SSI) is based on average earnings per start (adjusted for sex) per racing year. AEI favors nags who are sound and race often. SSI favors nags who race less often. Six of one, half a dozen of another. Both are useful indexes. As applied to sires (and later to broodmare sires), both are measures of “influence.” Both are based on a scale of 1.00 being average.

“Influence,” therefore, is a measure of ACTUAL RACING RESULTS. It can be quantified and standardized to a scale of 1.00 being average so that you are can see at a glance which sires (and broodmare sires) possess above- or below-average “influence.”

So if you can quantify the “influence” of sires and broodmare sires, why not do the same with sires of sires (P1 in the second generation)???? Why not do the same with sires of sires of sires (P1 in the third generation)???? Who not do the same with the other three positions in the third generation???? Why not do the same with the eight possible positions in the fourth generation????

These were all good questions, I thought, when I started asking them around 1988 or so. That is the year I joined Thoroughbred Times, then affiliated with Bloodstock Research Information Services (BRIS). BRIS loved to brag about its omniscient computer and all the “magic” it could perform. So I tried to interest them in this idea and persuade them to perform some “magic” for me.

They made one half-assed effort at it, using stakes winners only (NOT the correct way to go about it). They said they could not see anything “significant” in the numbers they generated.

They must have been BLIND as well as STUPID. I could see some definite patterns, even using stakes winners only. But I could not persuade them of that. This experience and other similar ones soured me on using their omniscient computer and all its “magic.” See To Build a Pile for a related story on this same theme.

Why was I so interested in quantifying the “influence” of sires from all portions of pedigrees in the first place???? Mainly because it was intellectually interesting, and I had a “thirst for knowledge” (as my high school chemistry teacher used to say). I thought that if I could quantify the “influence” of sires from all portions of pedigrees, I could come to a better understanding of how pedigrees really do and do NOT work.

“Uncle Joe” Estes was a major influence on me. I strongly recommend as required reading “How Joe Estes Transformed Breeding and Pedigrees,” Chapter 4 in Racehorse Breeding Theories. It is the longest and best chapter in the whole book (including my own chapter therein).

I felt like I was trying to extend the work of Estes myself. Now if you are familiar with Estes, you might wonder at this point if Estes would have approved of quantifying the “influence” of sires from all portions of pedigrees. Estes believed that sires and dams were by far the most important components of pedigrees. He might have seen me looking deeper into pedigrees as a distraction from sires and dams. He might have seen it as making names deeper in pedigrees appear more important that they really are.

Estes need not have worried. That was not my intention at all. After all, I was not trying to “sell” these numbers. If I had been trying to “sell” these numbers and make $$$$$ off them, then Estes should have been worried.

Reading between the lines, that is undoubtedly the reason why the powers that be at BRIS declined to explore the matter themselves. They could not find a market for it. They could not think of a way to make $$$$ off it.

So using the massive computer resources of BRIS was off the table. I had to think of ways to explore this subject on my own. I had to devise methodologies that did not NEED the computer to be calculated. All they needed was a lot of hard work, one pedigree at a time, and a fair amount of number crunching.  “Someone had to reach for the rising star/I guess it’s gonna be up to me.”

So I thought of some ways to explore this subject on my own and executed them over the years. I did so without massive computer resources to accumulate all the data and crunch it all for me. I did use the computer, but I used it merely to mine the data myself (one pedigree at a time) and then to crunch it myself. It was a lot more FUN that way.

Along the way I discovered that the best uses of these numbers were to discredit all kinds of assorted breeding theories (of which Estes would have thoroughly approved, the discrediting, that is). Many, many breeding theories rely on names in pedigrees still being “important” and capable of positive “influences” even from deep within pedigrees. Demonstrate that names in pedigrees are nothing more than names in pedigrees, and voila, many theories crumble into dust.

Geneticsts will tell you that the genetic inheritance of the Thoroughbred is pretty much a crapshoot. It is extremely unreliable and unpredictable (most breeding theories are dependent upon names in pedigrees being reliable and predictable, which they are NOT). I was able to demonstrate just how unreliable and unpredictable names in pedigrees really are.

Looking back on it, that is undoubtedly another reason BRIS did not want to explore this project with me. It undercut their own business, which was making $$$$ by selling pedigrees based on theories dependent on names in pedigrees being somehow reliable and predictable and mainly positive in “influence” (which they are NOT). I doubt if they understood this at the time, but they may have sensed it with their low, animal cunning (instinct for making $$$$$$).

So over the past 25 years or so I have amused myself enormously by mining tons of data one pedigree at a time and then crunching said data into iconoclastic conclusions. I think “Uncle Joe” Estes would definitely approve.

“The intellectual foundation of science is observation, logic, and skepticism.” I do not know if this is an actual quote from some famous luminary, but I found it on the internet and take it as my mantra.

Observation equals data mining and crunching. I have done tons of that over the past 25 years or so. I tried to extend the work of Estes in a logical fashion. In doing so I exercised plenty of skepticism regarding existing breeding theories. Some say I exercised a surfeit of skepticism. Better too much than too little when it comes to science.

In my next post I will try to explain what geneticists mean by “influence” (as opposed to how most people understand “influence”) and all the ramifications thereof.

This entry was posted in Uncategorized. Bookmark the permalink.

11 Responses to “Influence” (One Pedigree at a Time)

  1. Ned Williams says:

    Another fantastic post. Thank you! I have been searching for someone to extend Estes work and to confirm it. You have once again made the confusing clear and helped debunk some of the pedigree craziness. (If you want a great example of this pedigree backtracking, read the pedigree analysis on Mizdirection this weeks in the Bloodhorse) I had the feeling you were in the “Estes” Camp (hence my last post found under “Full of Sound and Fury”). The question that I am currently wrestling with is this: What are the chances of a graded stakes winning mare producing stakes winning foals? I have read the studies that contend that ones chances increase.
    However, I have never seen any data that speaks to this all important question: If a mare produces one stake winning foal, does that mare then have a better chance of producing a second (or third) stakes winning foal. Most buyers at sales look at the catalog page and use this premise: If the first dam produces one stake winner than they (the potential buyer) will pay more for the next foal (all else being equal) because the mare is a proven producer. Yet, my not so scientific study leads me to question this. Any thoughts? Thanks for your healthy skepticism and deep insight.

    • ddink55 says:

      Thank you very much.

      You seem to have posed two questions. The first is what are the chances of a graded stakes winner (mare) producing a stakes winner????

      The answer to that question was implicit two posts back. There were 2,052 foals among sales foal of 2003-2007 who were out of graded stakes winners. Those 2,052 foals produced 133 stakes winners (6.48%). I mentioned at the time that I thought that was a little low, but it is almost double the results for all other mares (2,276 stakes winners from 68,662 foals, or 3.31%).

      For the sake of argument, let us say that the average mare who is a graded winner produces eight foals. Therefore, her chances of producing at least one stakes winner are approximately eight times 6.48%, or close to 50%. Kinda interesting, heh???? Even graded stakes winners become producers of stakes winners only about half of the time. That is my estimate anyway. But that is approximately two times better than all other mares.

      On to the second question. If a mare has produced one stakes winner, does that make her more likely to produce another stakes winner (or multiple stakes winners) after the first one????

      If you have read your Estes, you should know that the answer to that is affirmative. I actually did a study of this back around 1991-1992 or so at TT. You could not possibly find it now. It appeared in one of their sales supplements, not in the magazine itself.

      It was based on about five years of the best yearling sales in the country. And yes, it was one of those that I did pretty much all by myself (not computer programmed).

      The basic conclusion of the study was that once a mare had produced a stakes winner, her value as a broodmare was just about the same as if she were a stakes winner herself. Prices were incorporated into this study as well, but there was very little difference in overall prices or results between mares who were stakes winners and mares who had already produced at least one stakes winner.

      There were of course some common-sense caveats to that. If a mare did not produce her first stakes winner until the age of 20 (as Mom’s Command did with Jonesboro, for example), that did not really help much if only because you might not even GET another foal out of that mare. Her one and only stakes winner might be her LAST foal.

      So as a general rule of course it was better if the mare got her first stakes winner EARLY rather than LATE. That was just common sense. If a mare got a stakes winner in her first foal, that was more impressive than a mare getting a stakes winner with her tenth foal (one out of one is better than one out of ten). Not to mention that the former had many MORE prospective foals in the future than the latter.

      As a general rule it was better if the mare got her first stakes winner in her first four foals than if she got her first stakes winner in her fifth foal or later. Just common sense, for the reasons cited above.

      The devil was in the details as usual. And I did break the numbers down as to whether the mare had her first stakes winner as her first foal, second foal, etc. The numbers were pretty clear. The earlier the mare got her first stakes winner, the higher the prices and the better the results for all her subsequent foals.

      Of course I can not PROVE all that. That is merely my recollection (which nevertheless is still pretty clear).

      I remember the late John M. S. Finney joking with me about this study. He related a story about it. One of his clients had read it and had a question: “What about the FIRST foals of stakes producers?????” We had a GOOD CHUCKLE over that. Anyway, hope this answers the question.

  2. Ned Williams says:

    Thanks for your explanation. You have helped me with some decisions regarding a program for a few mares going forward. This to me is the ultimate compliment of your work, when the theoretical is moved into real world application. I pulled out a dog eared copy of Estes this morning and will do a reread. (I am slow learner!!!) Thanks so much for the difficult number crunching and clear explanations that make your writing and thinking stand out.
    best,
    n

  3. vineyridge says:

    So what do you think of the European trend (at least in sport horse, not race) toward the use of breeding indexes which attempt to quantify heritability of traits for sires. Not sure about whether TB breeders in places like Germany use such indexes, but they are very, very common otherwise. For some interesting reading take a look, for example, at the KWPN website and the numbers that they generate for their approved stallions. It’s not just the Germans and Dutch, though, that are using this approach, but also the French and Swedes. What they’ve come up with seems to be computer driven breeding for heritable traits. Of course, for TBs, work would have to be done first to identify the conformational and other traits that contribute to racing success before the indexes could ever be developed.

    For example, if you have a mare who needs help with length of neck, you can search for a stallion who is reliably proved to pass on an acceptable length of neck. The breeding indexes are derived from all foals of the stallion and are VERY complicated computations, which include performance data for all the foals.. The indexes are similar to those used in cattle breeding to predict, say, the butterfat in milk or calving ease

    Do you think it’s likely that the TB industry will ever even consider doing something with breeding indexes?

    • ddink55 says:

      The only thing I know about such indexes is what I read in Chapter 3 of “Thoroughbred Breeding Theories” on the BLUP trotting index. I will quote its last three paragraphs:

      ” . . . but at its root, the BLUP was designed–and seems to function most effectively–as a means of assessing populations, not individuals. This is a fine state of affairs when breeders are testing thousands of chickens for volume of egg production from six to 12 months of age. Chickens, for instance, reproduce quickly, allowing relatively small improvements in selection to offer significant dividends over a short time.

      “However, applying a tool for population assessments to individuals can have some definite shortcomings.

      “In addition, horses are not cows, pigs, sheep, nor corn. They are not bred for a single trait or a pair of traits, and the expression of speed is not a purely genetic result. So many different systems go into the making and maintenance of an equine athlete that anyone wishing to use the BLUP index should take it as a tool, a starting point, or a means of gaining a broad view–not as an end in itself.”

      That seems like a fair assessment to me. To answer your last question, possible, but not likely.

  4. Ned Williams says:

    Can you point me in the right direction to find the actual chart for AEI dollar amounts by year. Russell Meerdink includes one (on page 90) in his compilation of Joseph Estes work, The Estes Formula for Breeding Stakes Winners. However, the chart only includes years until 1997. Any help would be greatly appreciated. I feel I have seen it somewhere, maybe on this blog. However, I am unable to locate it at present.
    As always, thanks.

    • ddink55 says:

      Sorry, can’t help. Apologies also for the delay in responding. Intended to do so Sunday morning. It was in the back of my head that there was SOMETHING I was supposed to do, but I could not remember what it was. Alas, those who possess this information (B-H, TJC), would rather charge you $$$$$ for this info than PUBLISH it for all the world to see. Apologies again.

      • Ned Williams says:

        No worries. Thanks so much for all of the difficult work that you do….for free no less!!!!
        I hope I speak for all your readers when I say: Thank you, Thank you, Thank you.

      • ddink55 says:

        Welcome. I forgot to inquire WHY you are in search of this information. It sounds like you want to calculate AEIs on your own in order to evaluate broodmares. If that is the case, I might suggest that you use SSI instead. Not because one is better than the other (both have their advocates), but SSI is available, and AEI is not for individual nags. To find the SSI of an individual nag (broodmare prospect or whatever), just go to BRIS and run a 105E (free race record and pedigree). The cutoff point for stakes-winning class is 4.0 for both AEI and SSI. Good luck!!!!!!!!!!

  5. Ned Williams says:

    You are right on target. Thanks. I’ll use the Bris SSI.
    best,
    n

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s