[This is a continuation of the second part of the deep analysis of pedigrees and performance in regard to the highly popular Rasmussen Factor for breeding. The remaining portions of this sometimes rather technical discussion will continue in coming days until this offering is complete.]
Results of the study
Of the 54,244 foals sold, 2,041 (3.76 percent) became stakes winners. That 3.76 percent is actually on the low side because of the number of foals sold more than once and hence counted more than once, as explained in part one. If the same foal were counted only once regardless of how many times it sold, that 3.76 percent would be more like 4.5 percent.
Of the 3,886 RF qualifiers sold, 144 (3.71 percent became stakes winners). The latter is obviously lower than the former. RF qualifiers underperformed unless it can be demonstrated that the quality of the stakes winners among the RF qualifiers was better than all the other stakes winners. How to go about doing so?
Perhaps a baseball analogy might be useful. Most people are familiar with batting averages and what they stand for and how they are derived. Basically it is the percentage of hits from at bats; 25 hits from 100 at bats is a batting average of .250, which is pretty close to the norm for the Major Leagues.
Percentage of stakes winners from foals among Thoroughbreds is similar to a batting average in baseball. The number of foals is the number of at bats. The number of stakes winners is the number of hits. The batting average norm is about .250, as I stated above. For the overall population of Thoroughbreds, about three of every 100 foals becomes a stakes winner (that would be a batting average of .003). Sales foals are somewhat
higher than the overall population. Good sires are even better, but very few sires left today manage to produce 10 percent stakes winners from foals (a batting average of .010).
Baseball also has a statistic called the slugging percentage. It differs from batting average in that it awards one point for each single, two points for each double, three points for each triple, and four points for each home run, then divides by the number of at bats.
I do not know what the norm for slugging percentage is, but I would guess around .400. I recently saw that the top slugging percentage in the Major Leagues this year was .706.
I thought it would be interesting to develop a slugging percentage to evaluate sires and/or groups of foals. Only two things are important and necessary to know in order to do so. What percentage of “hits” (stakes winners) did each group produce, and how good were those stakes winners (how many bases were they worth).
At first I simply decided to award each stakes winner with four points for each G1 win, three points for each G2 win, two points for each G3 win, and one point for each nongraded stakes win (all by black-type rules). That was very similar to the slugging percentage in baseball.
The problem with that rating system is that it underrates a stakes winner such as Luthier Fever (going back a few years, a foal of 1991), who earned $1,160,852 but won only one (restricted) stakes race. Conversely, it overrates a theoretical stakes winner who won six small (and probably restricted) stakes races but earned only $200,000. In the old system the former was rated a one and the latter a six.
So I decided to incorporate earnings into the new rating method. Specifically, I decided to give each SW 400 points for each G1 win, 300 points for each G2 win, 200 points for each G3 win, and 100 points for all other stakes wins, plus one point for each $1,000
earned. Now Luthier Fever is rated 1261 and the theoretical nag 800, which is clearly an improvement. Call these Performance Points (with 627 being the norm, or average for this particular group) and the rating derived therefrom the PPI (Performance Points Index).
Most people would say that the theoretical stakes winner is still overrated, but I want to give credit to horses who did race extensively and penalize horses who slink off to stud after very limited racing careers. In other words, this is a measure of actual accomplishment, not necessarily of ability or stud value. Most people complain about the lack of soundness in the modern racehorse. I decided to reward the stakes winners who did display soundness (raced extensively albeit more cheaply in most of the cases).
The only problem with using earnings is the number of such sales foals who went
over and won stakes races in Japan, where earnings are notoriously inflated (although they seem to be coming down to earth a little bit lately). I decided to deal with that problem by deflating Japanese earnings. I harked back to a study I did awhile ago which showed that North American-bred stakes winners in Japan earned about $2,350,000
on average, compared to $265,000 or so for all other North American-bred stakes winners in that time period. The ratio was about 8.85. I decided to be a little conservative and have deflated Japanese earnings by a factor of 8.5 in these ratings.
The biggest difference between slugging percentage and Performance Points is that the latter has a far wider range of possible results. For the former a hit is either a one, two, three, or four. For the latter the values assigned to stakes winners range from about 120
(horse who won one stakes race and earned about $20,000) to almost 10000. Pleasantly Perfect is the highest-rated stakes winner among these foals sold in 1999-2002. He won six stakes races, three G1 and three G2. That gives him 2100 points. He earned $7,789,880. That gives him 7790 points for an overall rating of 9890.
The 2,041 stakes winners among these 54,244 foals earned a total of 1,279,004 Performance Points. That is an average of 627 Performance Points per stakes winner. The 144 stakes winners among the RF qualifiers earned 89,368 Performance Points. That is an average of 621 Performance Points per stakes winner. The latter is slightly lower than the former. Therefore, the RF stakes winners were, if anything, of slightly lower quality than all other stakes winners in this group. Note that 7,931 of those 89,368 Performance Points for the RF stakes winners were earned by one horse, Ashado, who accounts for 8.9 percent of the total. The RF stakes winners would be a lot worse without that one horse, Ashado.
To review, the norm for the entire group was 3.76 percent stakes winners from foals. The RF group produced only 3.71 percent stakes winners from foals (98.67 percent of what it should have been). The norm for the entire group was 627 Performance Points per stakes winner. The RF stakes winners produced only 621 Performance Points per stakes winner (99 percent of what it should have been). Taking the two components together, 99 multiplied by 98.67 is 97.68, or .9768 on a scale of 1.00 being the norm. That is very close to the figure of 0.9754 quoted at the beginning.
The difference is due to rounding errors. The actual calculation I use is based upon Performance Points earned per foal. The entire group of 54,244 foals earned a total of 1,279,004 Performance Points. That is an average of 23.58 Performance Points per foal. The 3,886 RF qualifiers earned a total of 89,368 Performance Points, an average of 23.0 Performance Points per foal. 23.0 divided by 23.58 is 0.9754. Whichever way you calculate it, you come to the same conclusion, that the RF qualifiers in this group had fewer stakes winners per foal than the overall group and stakes winners of lower quality than the overall group, albeit both by tiny margins.