Latest article from Jason Hall on ATR

[ides of ice]

He's addressing Stallion durability,says Pollards Vision is the worst...

http://www.thoroughbredreview.com/

[kimberley mine]

I smell cow chips.

Look at the third to last column--average starts per racing season. If he's using that number at all, he's getting skewed results, because that figure all by itself is misleading. Look at Giant's Causeway. His average starts per year of racing is 6.5.....he made 3 starts at 2 and 10 starts at 3. Except for the BC, all of his racing was in Europe, where the running season is much shorter than the US racing calendar, where horses run all year round. Big-time Euro racing is from April to October. If you compare Giant's Causeway and Grand Slam, is GC necessarily less sound than Grand Slam because Grand Slam ran two more times at two, and equally as many times at three?

For that matter, is Pleasantly Perfect less sound than Giant's Causeway? His numbers are pretty skewed by his 3yo year, in which he only raced once..and the average starts/season calculation says he ran for 4 seasons, when the horse never raced at 2. Then there's Leroidesanimaux, who as a Brazilian-bred lost a full year in age by jockey club reckoning when he came to the northern hemisphere.

And that's just the one column.

Next up: percent starters from foals. For stallions with at least 4 crops older than 3, it's a useful metric....sometimes. North Light jumps out here with only 60% starters from foals. Is it because they're too fragile to run, or is it because they are too slow? Given how poor a job he's done so far--median earnings a paltry $6900, compared to fellow turf champion and freshman of 2006 Kitten's Joy with a median earnings of $23,000--the idea that North Light foals are so much more brittle than Kitten's Joy foals that 25% fewer of his foals ever make it to the track is laughable. Money matters. There are many more apt comparisons there--both KJ and NL are stallions supported by their owners' racing operations; both the Ramseys and Stronach race a lot of their homebreds. If you're a racer-breeder, you don't waste your training money on a horse that is too slow to win.

He says he's controlling for racing class, but is he controlling for mare power? Look at Arch and Flatter here, versus Point Given and Giant's Causeway. Both Arch and Flatter started with mostly blue-collar mares and worked their way up the ladder. Claiborne occasionally gave Arch in particular a good one, but nothing like the books of mares that PG and GC got. Race breeders like the ones patronising Arch and Flatter at $5000 are going in with cheaper mares BUT the expectation that they run more often, e.g. that they'll race to make money. Then if they hit and you get a high percentage of GSW, you get better mares, meaning more GSW. Thus, just controlling for GSW percentage isn't sufficient to overcome the potential soundness bias that racing mares bring in.

[kimberley mine]

I thought about this a bit more and it's even sloppier statistics on second thought than it was on first blush. Here are some stats I did a couple of years ago:

http://www.pedigreequery.com/forum/view ... e+starters

It was after a series of arguments/discussions about whether low numbers of starts correlated to low soundness. There was no correlation between number of starts and confirmed on the track breakdown (in a couple of cases, known injury in a workout), but there was a very strong correlation between number of starts and racing class, and not just stakes-level racing class.

Here's a short version of the findings:

Nonwinners averaged 6.23 starts, median 5.5.
1-3x winners averaged 15.68 starts, median 15.
4-6x winners averaged 32.28 starts, median 33.
7-9x winners averaged 42.43 starts, median 35.
9+x winners averaged 53 starts, median 61.
Blacktype winners averaged 25.18 starts, median 17.

In the money (ITM) is defined as win/place/show. NITM = never ITM.

EDIT--NITM (12.5% of runners) averaged 3.73 starts, median 2.
ITM 1-5x (37.21%) averaged 10.35 starts, median 10.
ITM 6-10x (24.81%) averaged 19.59 starts, median 17.5.
ITM 11-15x (10.08%) averaged 32.38 starts, median 31.5.
ITM 16-20x (6.98%) averaged 35.56 starts, median 35.
ITM 21-25x (6.98%) averaged 45 starts, median 45.
ITM 26+x (3.88%) averaged 66.2 starts, median 63.

Horses that were never in the money, not even once, barely started at all, while even professional maidens (always the bridesmaid) usually ran 17 times (median for that group) which is close to the industry average.

There's no way on this green earth that a table of information saying percentage runners to starters and percentage graded stakes winners can control for racing class of the horse--which, as my little pocket study showed, has a tremendous amount of influence on how often a horse runs.

That's where we come back to North Light. He has a poor starters-to-foals ratio, but then when you see that barely half of his starters ever win, the paltry median earnings, and where 5% of his starters (2% of his foals of racing age) account for 54% of his total progeny earnings.....where his foals on average run 9 times and on average are in the money 31% of the time and STILL have a median earnings of under $7000.....they're just too slow to be economical to race. They could be the soundest horses alive, and if they're too slow to run, they don't run.

As for Pollard's Vision, his track values are heavily skewed by Blind Luck. Toss her out, and his average earnings per runner drops by almost $20,000 to 40,424. Throw out Blind Luck, and the average earnings per start drops from 4527 to 2923--precipitious drop, and implies a lot lower racing class. That's borne out in the low median earnings of $15k, half of his foals will make less than that on the track.

Long story short, without controlling for the influence of the dams and for the racing class of the resulting offspring, his conclusions are somewhere on the spectrum between waffle and baloney. I wouldn't put much faith in them.

[LB]

There's no way on this green earth that a table of information saying percentage runners to starters and percentage graded stakes winners can control for racing class of the horse--which, as my little pocket study showed, has a tremendous amount of influence on how often a horse runs.

Long story short, without controlling for the influence of the dams and for the racing class of the resulting offspring, his conclusions are somewhere on the spectrum between waffle and baloney. I wouldn't put much faith in them.

^^ This. Well said.

[Barcaldine]

Just another example of a drive-through pedigree "expert" peddling misinformation in the hope that buyers aren't all like Kimberley Mine.

[kimberley mine]

Just another example of a drive-through pedigree "expert" peddling misinformation in the hope that buyers aren't all like Kimberley Mine.

Something else from that little thing I did 2 years ago....is it few starts leading to lack of wins, or lack of wins leading to few starts is a valid question. What I found in that sample group is that no horse in the group ran more than 12 times after its last in the money (w/p/s) performance.

I keep pounding on North Light, but he is another reason why using the ratio of runners to starters is inherently and fatally flawed. North Light was a router, out of a router whose greatest victory was the 4000m Prix du Cadran and whose sire was notorious for siring routers (Rainbow Quest, sire of 2 St Leger winners, an Ascot Gold Cup winner, a San Juan Capistrano winner, a Prix Royal-Oak winner, etc). This horse is bred to stay all day on soft turf, he ran like he wanted to run all day on soft turf, and his best horse is any indication, an Irish St Leger winner, he's siring horses who want to run all day on soft turf. Despite this, he's at stud in a place that wants horses who run on dirt for a mile or less. Mismatch much?

There is precisely ONE graded stake over a mile and a half in North America: the San Juan Capistrano. There are proportionally extremely few races over a mile and a half anywhere in the continent and at some tracks you are hard pressed to find races at a mile. Where in the world are the North Lights supposed to run? Even at Woodbine on the synth or turf, a horse who needs 2 miles isn't going to find a race that suits.

But you watch, with his 60% starters from foals and abysmally low proportion of graded stakes winners, he's going to get a totally unjustified F for soundness.

If you can't win them with facts, bludgeon them with bullssss.

[Barcaldine]

"If you can't win them with facts, bludgeon them with bullssss."

Hall, Rogers, Porter, Fernando and a few other legends in their own minds are certainly proficient in the latter.

[Sysonby]

I know its fun to play with numbers but sometimes you have to roll up your shirtsleeves and review a stallions record using the wonderful free resources at our disposal (especially Equibase and the Stallion Register). I did that recently for a stallion who intrigued me and looked at the both the most accomplished runners and his most expensive at the sales. By the time I got to the fifth "ran two times and then disappeared for over six months" runner, I threw him off my list. I also noticed that barring first time starters, he had nothing notable in the pipeline. But that analysis alone for one stallion took about three hours.

Numbers are great but I think you need to know what's actually going on which still means research.

[jellac]

Just as there are 'horses for courses' there are 'stallions for different markets/purposes' -

Using ATR's list of values used I noted longtime top stallion in Texas, Valid Expectations line of figs just under the highly regarded Unbridled Song's - quite a difference which reflects both the regional differences in where these stallions stand, their completley different conformational type and pedigree, the HUGE difference in what they 'throw' to their offspring and the balance between the risk of their widely different stud fees and their offspring's market potential....then you can throw in the vast difference in their mare books. It's apples and oranges folks! So of course they look widely disparate in the numbers. Both are good, even excellent stallions for their respective 'markets' but the resemblance ends there. That's one of the inherent problems with trying to compare so many different individuals - and while statistics can 'smooth' out minor differences given a sufficient sample to begin with it can't - AND SHOULDN'T turn an "apple" into an "orange" for simplicity of comparison's sake.

Goodwork Kimberly -

[Sysonby]

I think the other thing to keep in mind is that when we select a stallion, we are not buying a commodity off a shelf but trying to project how that stallion will match with a specific mare. I believe that "soundness" isn't this amorphous thing but usually a function of conformation. A stallion that tends to back at the knee may not have the sturdiest individuals and the best starters ratio all other things being equal.

This is where advisors can be useful. It is not sheer loonyness as to why some stallions have better runners than sales horses. Since I am always looking for "value" those are the waters I swim in. More often than not the stallions with great racing numbers but low fees have low sales figures too and tend to also have a bunch of hammer headed, back at the knee, small, offset individuals that get punished in the sales ring. Sure you can get the one that punches a hole in the wind but statistics being what they are, you can also get the one that can't sell and can't run. Yippee! So the market requires the risk to be built in the fee discount.

[kimberley mine]

Just as there are 'horses for courses' there are 'stallions for different markets/purposes' -

Thanks.

The one and only Valid Expectations horse I have met looks like he should have skipped the track and gone to the sorting pen, and I mean that as a compliment. He was a big, sturdy horse who looked like he was built to work and who wouldn't think twice about lugging around a big sturdy cowboy and a big sturdy saddle for 8 hours a day.

I think the only way to do a comparison on whether a stallion's get race more or less than a national average is to start with ones with at least 500 foals older than 3, and then break that group into subgroups: horses who never went into training, horses who were trained but never went to the track, horses who never won, horses who won at stakes level, horses who won at allowance level, horses at a high claiming but below allowance level, and horses at low claiming levels. You would need to look at frequency of races AND control for opportunity.....CLH on this board in the last week has lamented the lack of allowance races at a mile within a 3-state area for her Invasor colt, and she doesn't want to drop him and risk losing him in a claim. ...and as Sysonby said, you would need to investigate whether the horse had any long breaks in training that weren't explained by outside factors, such as Woodbine being shut down for the winter, or an entire meet being cancelled (was that Hollywood a few years back, the turf course?). And within each group, you would have to see if the results were different between sexes--females, entire males, and castrated males.

After you are done with THAT, then if you wanted to compare two or more different stallions, then you get into the wonderful world of multivariate distributions and that's an entirely different kettle of fish.

It's one thing to do what Sysonby did, to do a simple results-based evaluation of what a given stallion was throwing, and making an educated decision based on that. It's entirely another to categorically state based on cursory examination of incomplete information that a stallion has unsound foals, and publish it as an advertising gimmick.

I believe that "soundness: isn't this amorphous thing but usually a function of conformation. A stallion that tends to back at the knee may not have the sturdiest individuals and the best starters ratio all other things being equal.

And this comes back to Jellac's point, that some stallions are better suited for certain racing environments than others. Back at the knee isn't the conformational kiss of death on a good/dead turf track that it is on fast dirt. Crooked pretzel legs where the horse moves through it well isn't the same as crooked pretzel legs where the horse runs and you're afraid that the leg will come unattached at the shoulder. And a bad farrier can mess up even the soundest horses, quickly.