This is part II in my thoughts about broodmare sire, percentage of stakes winners; as I'm posting both parts this morning, you might want to start with the earlier one....In looking at broodmare sire stats at the bloodhorse.com site, there is one disadvantage, which is that I am looking at the 2011 top 100 broodmares sires by earnings. My interest is in finding truly good and truly bad broodmare sires, and if the broodmare sires are bad enough, perhaps they're not on the list at all. That said, the top 100 is not a small list; being in the top 100 is nothing to write home about, and signifies a large number of runners, more than anything else. So I do think there are some underperforming broodmare sires on there.
My first candidate for that was Gulch. He is #52 on the earnings list, but his top earner is Glenwood Canyon, with just $164,907. More shockingly, he has just two 2011 stakes winners. Of the 51 broodmare sires ahead of him in earnings, the lowest number of 2011 stakes winners is just 5. Twenty-five of the top 28 earning broodmare sires have double-digit stakes winners in 2011. On the whole top 100 list, all broodmare sires but Gulch, Bertrando, and Regal Classic have more than two 2011 stakes winners, and Bertrando and Regal Classic have fewer starters than Gulch.
It then occurred to me that it was stupid to be evaluating Gulch by his 2011 number of stakes winners. What does that mean, really? Why should I invite flukes into the process? Gulch's overall percentage of stakes winners as a broodmare sire is 3.6, with 1646 foals. The percentage is on the low side, but he had plenty of company on the top 100 list with it. Gulch's percentage of stakes winners is not remarkably bad, the way it has been in 2011.
This whole line of thinking about 2011 versus the duration of the stallion's broodmare sire history got me thinking about something my father and I used to argue about before he passed away. He was always scouring stallion statistics year by year, looking for stallions who had had drop offs, who had gotten old. Actually, he believed young stallions could drop off, too, because he believed mutations were the driving force to this. I explained to him that the degree of expected fluctuation because of sample size was so great that a mutation could never be inferred, and career statistics were far more helpful in predicting future performance than recent statistics.
Now, it occurred to me that year by year broodmare sire statistics are the perfect place to point out the random component, outside of mutation. The daughters of a stallion are an ever changing lot, bred to an ever-changing lot of stallions. If I could show that broodmare sire stats fluctuate as much year to year as stallion stats, wouldn't that be instructive?
I knew that Gulch had had 60 stakes winners overall as a broodmare sire, but just two in 2011. That got me wondering what his number of stakes winners were in other years, and whether the role of random fluctuation could be well demonstrated with him. Here are his number of stakes winners and starters yearly since he began appearing on the top 100 broodmare sire list in 2004.
2011 2 389
2010 8 388
2009 4 374
2008 4 351
2007 9 313
2006 6 257
2005 7 167
2004 6 257
I think the numbers make my point, but not as lucidly as I would have hoped. Gulch's average number of stakes winners from 2004-2010 was 6.3 with 301 starters, compared to 2 with 389 starters this year. His precentages in 2008-2011 have fallen off from the 2004-2007 time. Whether the main factor is coincidence, or a slight aging on teh average and subsequent decreased effectiveness of his broodmares, or something else, I have no way of knowing right now.
No comments:
Post a Comment