Some of you may remember the dialogues between “Uncle” Kent Hollingsworth (late editor of the Bland-Horse) and Col. Ellsworth. A friend and I had one of those similarly foolosophical conversations a few months ago.
It started off with the folly of thinking you got a “bargain” because you bought a progeny of a “good” stallion “cheap” at public auction. A Distorted Humor for $10,000, for example. My friend and I agreed that such purchases were not likely to be “bargains” at all.
At the same time we agreed that the most-expensive progeny of the same stallions might be overpriced. My friend was even more particular. He thought that the best bargains were to be found in the upper-middle tiers for each stallion, say from approximately the 50th to 75th-80th percentiles.
I quickly came up with three good examples confirming that notion from my recent work with sales foals of 2003-2007: Zenyatta, Curlin, and English Channel.
Zenyatta sold for $60,000 as a yearling in 2005. The 42 yearlings by her sire (Street Cry) sold (in North American only) for a median of $53,500 in that same year.
Curlin sold for $57,000 as a yearling in 2005. The 51 yearlings by his sire (Smart Strike) sold for a median of $55,000 in that same year.
English Channel sold for $50,000 as a yearling in 2003. The 29 yearlings by his sire (Smart Strike) sold for a median of $47,000 in that same year.
All three sold for prices just above the medians for their sires that year. All three would have been categorized as “upper middle” in terms of prices relative to the progeny of the same sires in the same year.
It just so happens that Zenyatta, Curlin, and English Channel were by far the best stakes winners among all sales foals of 2003-2007. With all three being categorized as middle priced, it seemed evident to me that the middle price group would fare well against lower and upper price groups.
So I told my friend that I would work on this project when I got a chance. First we needed to establish the parameters. It was decided to examine only yearlings sold (not weanlings nor two-year-olds). It was decided to examine only sires with ten or more yearlings sold in a given year (obviously because we needed a wide basis for comparison of prices).
It was decided to examine only sires with a median of at least $10,000 for a given year. Also that those sires had to have had at least one yearling sold for at least $100,000 in that year. So we were looking for commercially viable sires with a fair number of yearlings sold in a given year.
Next we had to decide how to define the three groups. It was decided that low would be anything sold for less than the median for that sire for that year. Call that Group A.
Middle would be defined as anything sold from the median to twice the median inclusive. Call that Group B.
High would be defined as anything sold for more than twice the median. Call that group C.
By definition the lowest possible median was $10,000. So in that case less than $10,000 would be Group A (low), $10,000 to $20,000 would be Group B (middle), and $20,001+ would be Group C (high).
Some of the medians were quite high. Storm Cat, for example, had a yearling median of $825,000 in 2006. In that case less than $825,000 was classified as Group A (low), $825,000-$1,650,000 as Group B (middle), and $1,650,001+ as Group C ( high).
I found the time to work on this project, and voila, it is done. Here are the results.
Of the 70,714 sales foals of 2003-2007, 21,944 qualified (about 31% of the total). They divided into the three groups as follows:
A (low) 10,432 47.5%
B (middle) 6,474 29.5%
C (high) 5,038 23%
This split was deemed acceptable. Group A (low) was less than 50% because it was designated as LESS than the median. All prices that were EXACTLY the median fell into Group B (middle).
Here are the prices for the three groups:
Group Average Maverage Price Index
A $32,777 147.99 0.907
B $100,071 282.60 1.73
C $244,715 419.47 2.57
Totals $101,288 250.03 1.53
Here are the racing results for the three groups:
Group Foals SWs % APPPSW PPI Price Index
A 10,432 290 2.78 586 0.80 0.91
B 6,474 345 5.33 748 1.97 1.73
C 5,038 397 7.88 802 3.12 2.57
Totals 21,944 1,032 4.70 723 1.68 1.53
APPPSW stands for average Performance Points per stakes winner, my measure of the quality of the stakes winners involved, 603 being average.
The group as a whole is pretty good, with a Price Index of 1.53 and a PPI (result) of 1.68. They had prices about 53% above average and results about 68% above average.
Group A had poor results, with a Price Index of 0.91 and a PPI (result) of 0.80. They had prices about 9% below average and results about 20% below average. So we were right about that much anyway.
Group B had good results, with a Price Index of 1.73 and a PPI (result) of 1.97. They had prices about 73% above average and results about 97% above average. We were correct in thinking that B would have good results.
The only fly in the ointment is that the Group C was even better than Group B, allowing for its higher prices. Group C had a Price Index of 2.57 and a PPI (result) of 3.12. They had prices about 157% above average and results about 212% above average.
Group B improved about 14% on its prices (1.97 divided by 173 is 1.14). Group C improved about 21% on its prices (3.12 divided by 2.57 is 1.21). That is my reasoning behind the assertion that C was better than B, even allowing for its higher prices.
All of the above is comparing these 21,944 foals against the results for all 70,714 foals. We can also compare the 21,944 foals with themselves, assigning them prices and results of 1.00 and going from there. If you look at it that way, the results are as follows:
Group PPI Price Index
A 0.48 0.59
B 1.17 1.13
C 1.86 1.68
Totals 1.00 1.00
Nothing much changes except that the Group B now looks much closer to its prices than before (1.17 to 1.13). Group A is still poor (0.48 to 0.59), and Group C is still good (1.86 to 1.68).
I must admit that I found some of these results pretty surprising. Both my friend and I thought that B would be better than C, allowing for prices. We thought that Zenyatta, Curlin, and English Channel would put the B group way over the top. Instead, it appears that those three are the only reasons B is any good at all.
I will use only Zenyatta and Curlin as an example. If you subtract Zenyatta and Curlin from the B results, its PPI diminishes from 1.97 to 1.76. That is just barely above its Price Index of 1.73. If you compare the 21,944 foals to themselves and do the same thing, B has a PPI of 1.05 without Zenyatta and Curlin, which is actually below its price Index of 1.13. That tends to substantiate my statement that the only reasons B is any good at all are named Zenyatta and Curlin.
I was really surprised that C had better stakes winners overall than B (802 to 748 average Performance Points). With Zenyatta, Curlin, and English Channel, I thought B would be a lot higher in this category. I was still somewhat puzzled as to how C had better stakes winners than B despite those three.
So I checked out the percentage of stakes winners with 2,000+ Performance Points from foals for each group. A had 13 such stakes winners from 10,432 foals (0.125%). B had 18 such stakes winners from 6,474 foals (0.278%). C had 30 such stakes winners from 5,038 foals (0.595%). Mystery solved. C just consistently had more of the better stakes winners (those with 2,000+ Performance Points).
I can think of two morals with which to conclude this post. The first is that if you have a high enough number of foals, individual stakes winners (even those as good as Zenyatta and Curlin) do not skew the results all that much. The second is that while it is OK to speculate and make guesses as to what the results will be, there is no substitute for doing the actual data digging and number crunching.
I would have loved to show that Group C was overpriced, because I think that many of these sales foals were ridiculously overpriced. But alas, my own numbers confuted me. C’est la vie.
I will also post a sidebar on a related subject (the mathematics involved). Look for that as well if you are so inclined.