Horse running through field

Exception Proves the Rule

Just to clear up two questions that might arise from my last post on Galileo’s conditions of possible effectiveness (let’s call it CPE) with mares that have Sharpen Up in their ancestries.

First, that post established that Galileo’s probability of siring superior runners out of mares that descend from Sharpen Up through either Kris or Diesis is at least .28. Doubly Sure, the dam of Kris and Diesis, seems to be the key factor, especially considering that Cape Blanco (G1) is out of a daughter of Doubly Sure sired by Presidium, by General Assembly rather than Sharpen Up. That leaves open the question how well Galileo might handle mares in descent of Doubly Sure through her daughters, and I would have mentioned it except that Galileo has had only one such mare.

Second, I mentioned that Galileo had sired a superior runner out of only one mare of the 29 that descended in some way from Sharpen Up, but not through Kris or Diesis. So, the problem is, if descent through Kris or Diesis is a CPE governing Galileo’s relation to Sharpen Up, how can 16.5-f Marathon S. (Sandown, 2010) winner King of Wands (Galileo ex Maid to Treasure, by Rainbow Quest and an image in the Tarot, I believe) be explained? His third dam is by Sharpen Up.

The answer is that, even if a mare fails to fulfill a CPE relating to one of her ancestors, she may still fulfill other CPEs that control a sire’s potential. Now, Galileo doesn’t have a great strike rate with Rainbow Quest–only 4/24–and his record with Blushing Groom is 9/75. What is more likely decisive in the ancestry of King of Wands’ dam is her broodmare sire Indian Ridge, with which Galileo has a strike rate of 4/11. Moreover, his overall record with Indian Ridge’s sire Ahonoora is 6/18. That’s a superior-runner probability of at least .33.

It’s clear, then, that Galileo’s CPE relating to Indian Ridge and Ahonoora, whatever it might be, is far less stringent than his CPE relating to Sharpen Up. Furthermore, because of Sharpen Up’s generational distance in this case, his normal negative effect on foals by Galileo (when not descending through Kris or Diesis) is much less likely to be expressed than the highly favorable effect associated with Indian Ridge.

Discovering the conditions of possible effectiveness is a critical function of pedigree interpretation, and, while it can only yield probabilities comparable to those of, say, short-term weather forecasts, those probabilities are a lot more reliable than reading tea leaves, questioning the Ouija, or trusting in the make-believe deficit reduction in Paul Ryan’s budget proposal.

UPDATE: In the Tarot, the King of Wands denotes the ability to persevere toward a goal, overcoming all obstacles by exceptional means. Seems appropriately named.

The Possible and the Probable

by Roger Lyons

In previous posts I’ve noted that breeding methods are subject to conditions of possible effectiveness. That is to say, any given topic of pedigree discussion–a nick, a method of inbreeding, any combination of one ancestor with another–can yield a stakes winner if certain conditions are met by the specific cross. By definition, a cross that does not meet any conditions of possible effectiveness has virtually no chance of yielding superior performance. We can’t always know whether or not a cross meets conditions of possible effectiveness, but, given what is at stake, any evidence of such conditions is worth considering.

I would go so far as to submit that the conditions of possible effectiveness are crucial to determining the probability of superior performance from a cross. Here’s a case in point.

Last year I mentioned the nick between Galileo and the sons of Doubly Sure–Kris (by Sharpen Up), Diesis (by Sharpen Up), and Presidium (by General Assembly). At that time, 25 mares with one of those three sires in their ancestries had produced foals by Galileo, and six different crosses had yielded superior runners, including three G1 winners (Sixties Icon, Lush Lashes, and Cape Blanco) and New Zealand, which ran second in the Irish St. Leger (G1). Since then, Gallic Star, whose third dam is by Kris, won the listed Silver Tankard S. in England, bringing the total number of superior runners to seven.

Thus, assuming a mare of similar quality in some descent of Kris, Diesis, or Presidium–all out of the mare Doubly Sure–the probability that she can produce a Galileo foal capable of superior performance is at least .28, the probability of a G1 winner at least .12. Don’t get caught up in the fact that Kris and Diesis are by Sharpen Up. Galileo’s record of 7/53 with Sharpen Up suggests only a .13 probability of superior performance (not adjusted for the conditions of possible effectiveness).

But, we know from the numbers that the presence of a son of Doubly Sure in the ancestry of a mare is a condition of possible effectiveness for offspring of Galileo. That condition is met by 24 of the 53 mares with Sharpen Up in their ancestries, and six of those seven dams of superior runners met that condition. This means that Galileo’s strike rate with mares in descent of Sharpen Up, but not through a son of Doubly Sure, was only 1/29, with a probability of only .034 of yielding a superior runner. In other words, those cases do not meet Galileo’s conditions of possible effectiveness with Sharpen Up.

It’s perfectly reasonable to think of the probability of success as a frequency of superior performance from opportunity, but that reasoning goes wrong in thinking of opportunity as nothing more than the number of times a cross has been tried. It might be fair to say that the number of times a cross has been tried constitutes its nominal opportunity, but the real opportunity is defined by the cases that actually meet conditions of possible effectiveness. An “opportunity” that has virtually no chance of success is an opportunity you can do without, especially when trying to pin down a true probability of superior performance.

Pedigree, Conformation, and Zenyatta

by Roger Lyons

My last post explains why a recent post by Frank Mitchell gave me second thoughts about the appropriateness of Giant’s Causeway as a mate for Zenyatta in spite of the sterling statistical profile he has with her ancestry. My underlying point is that what’s on paper can and must be interpreted in light of what’s on the ground–and vice versa. That point drew some thoughtful comments bearing substantively on the case and more generally on the relation between pedigree and conformation.

First, Frank elaborates his reasons for thinking Giant’s Causeway might not be quite right for her, and it has a lot to do with her broodmare sire, Kris S. Frank not only casts his gaze on a lot of horses, but he also measures them, so you can be sure that he’s not speaking from casual observation. When he casts his gaze on Zenyatta, he sees a lot of Kris S. and some Troy, the broodmare sire of her sire, Street Cry.

Then Michele shares the experience, as a breeder, of having tried both approaches–breeding largely on pedigree and, alternatively, breeding largely on conformation, concluding that neither approach seems to make much difference in the frequency of favorable outcomes. Michele’s experiment was not conducted in a lab. It unfolded at much cost and over many years of trying to breed the best horses possible, and, as you read the comment, you get the sense that it rings true.

In the last comment so far, Greg correctly concludes that the distinction between pedigree and conformation is nothing more than a matter of emphasis. After all, he explains, statistical analyses that assess how a given stallion has done with mares representing a given ancestor actually do capture conformation issues–although indirectly. If you read Frank’s “The Weekender Pedigree” (and who doesn’t?) at The Paulick Report, then you know how much he’s into pedigree even though his science is biomechanics. The opposition routinely invoked by the cliche “pedigree vs. conformation” exists only because we associate pedigree analysis with one broad category of facts and conformation with another.

Greg proposes marriage of the two approaches, and he’s right. When Frank says that Kris S. is the major player in Zenyatta’s conformation (note the implication that pedigree and conformation are inseparable), it relieves a lot of statistical pressure. Rather than assuming Zenyatta’s entire ancestry to be more or less uniformly relevant, the focus can shift to Kris S.

Here are the candidates, along with their numbers with mares representing Kris S.: A.P. Indy (2/6), Galileo (1/2 and Frank’s choice), Giant’s Causeway (3/10), Invincible Spirit (1/2), Lemon Drop Kid (1/3), Mineshaft (2/5), Oasis Dream (0/2 with Kris S., but 7/35 with Roberto and great supporting numbers), Speightstown (0/2 with Kris S., but 2/10 with Roberto and good supporting numbers). And, by the way, Songandaprayer is 2/4 with Kris S. and may be a better choice for Zenyatta than better stallions that have poor or questionable numbers with Kris S.

Zenyatta Plus ?

by Roger Lyons

As an earnest reader of Frank Mitchell’s blog, I attended with interest to his remarks about a mate for Zenyatta, especially his misgivings about Giant’s Causeway as a possible match. Now, Frank is an expert in biomechanics, so I’m supposing that, when he says, “I don’t especially like him for this mare,” he means he sees a physical mismatch somewhere along the contours of the two individuals. When Frank speaks, I listen, but I don’t much like what I hear in this case.

That’s because on paper, which is where my life unfolds in this business most of the time, Giant’s Causeway is the best match out there among proven stallions. Don’t get me wrong. Frank’s choice is Galileo, and he also has a great profile on paper, but it’s not as factually confirmable as that of Giant’s Causeway–on paper.

Just for reference, Zenyatta is by Street Cry, by Machiavellian and out of Helen Street, by Troy, and her first, second, and third dams are by Kris S., Forli, and Hoist the Flag, respectively. Giant’s Causeway has had no opportunity with mares by Street Cry (hardly any stallion has), but he has a superior-runner strike rate of 2/7 with Machiavellian and 2/7 with Troy. He’s also 1/2 with Helen Street, the dam of Street Cry, because she’s also the second dam of four-time G1-winner Shamardal, by Giant’s Causeway.

On the other side of her pedigree, Giant’s Causeway has a strike rate of 3/10 with Kris S., 10/117 with Forli (a bit weak, admittedly), and 8/46 with Hoist the Flag. The way I add it up, Zenyatta’s ancestry scores in the 94th percentile of all mares that have been bred to Giant’s Causeway.

I’m not going to get in a fight with Frank over this because I know when to back down. The truth is–as much an embarrassment as it might be to those of us who specialize in pedigree–what’s on the ground has the right to veto what’s on paper. So, I’m going to defer to Frank on this and back Galileo although, in deference to what might actually happen, A.P. Indy has a very good profile, too. The Mr. Prospector-line stallions that have been suggested–not so much.

Galileo and the Sons of Doubly Sure

by Roger Lyons

The cross that yielded Cape Blanco (Galileo-Laurel Delight, by Presidium, by General Assembly), recent winner of the Dante S. at York, is formally an A nick based on the Sadler’s Wells-General Assembly cross. Galileo may never get enough opportunity with Presidium mares to confirm a simple sire-broodmare sire nick in the conventional way, but this is not a conventional nick.

The fact is that Galileo has done a lot better with mares whose ancestries contain Presidium’s dam, Doubly Sure, than he’s done with Secretariat-line mares. Seven of his 27 mates that had Doubly Sure in their ancestries produced superior runners by him, including Cape Blanco’s dam.

Now, Doubly Sure is also the dam of Kris and Diesis, both by Sharpen Up, and they both have simple A+ nicks with Galileo. Furthermore, the numbers say Galileo’s nicks with Kris and Diesis are not attributable to their sire line. Galileo has sired foals out of 53 mares that have Sharpen Up in their ancestries (the whole ancestry, not just the sire line), and the only mares among them that produced superior runners by Galileo are those that have Sharpen Up descending through either Kris or Diesis. None of the 27 mares that have Sharpen Up through strains other than Kris or Diesis produced a superior runner by Galileo. Yet, Galileo’s combined strike rate with Kris and Diesis is 6/24 (25% superior runners).

Overwhelmingly, the numbers point to Doubly Sure as the key factor, and that’s the proof that Cape Blanco is bred from a simple nick with her son, Presidium.