Ticker
  • In his first race since the Travers, Destin finishes 3rd in the 8th at Belmont.Posted 1 hour ago
  • The well bred Souper Tapit (9-5) runs to the money in the Marine Stakes.Posted 1 hour ago
  • Cloud Computing will not run in the Belmont Stakes, instead waiting for Saratoga's Jim Dandy.Posted 2 hours ago
  • Cupid (7-1) makes a winning comeback in the Grade 1 Gold Cup at Santa Anita.Posted 23 hours ago
  • Illuminant (8-5) sneaks up the rail to capture the Grade 2 Monrovia.Posted 23 hours ago
  • Cowboy Culture (5-2) is the clear victor in the Grade 3 Arlington Classic.Posted 1 day ago
  • Lady Eli (3-5) keeps Goodyearforroses at bay to win the Grade 1 Gamely.Posted 1 day ago
  • Finley'sluckycharm (1-5) runs away with the Grade 3 Winning Colors.Posted 1 day ago
  • Stormy Liberal (1-2) continues downhill turf success in the Grade 3 Daytona.Posted 1 day ago
  • Starship Jubilee (10-1) rallies to Grade 2 Nassau win.Posted 1 day ago
Breeders Cup 2015
Belmont Stakes 2017


HRN Original Blog:
ThoroughMetrics's Blog

Year to Year Correlation of Sires' Two Year Old Offspring

Animal Kingdom GP 615 X 400
Photo: Eclipse Sportswire

One of the difficulties of analyzing thoroughbred pedigrees is that (unless you have a substantial budget to purchase data), the information needed to test the predictive value of various statistics isn’t available.  Unfortunately, if any of those who do have access to the necessary data have done any studies on how well the various statistics work, the results have seem to have been kept private.

 

The most widely know measure of stallions’ success are all earnings based.  These include ranking them purely by earnings, and measures such as AEI which are derived from earnings per start or earnings per starter.  As far as I know, nobody has ever published analysis of the predictive value of these statistics.  There’s an appearance of consistency (which would indicate some predictive value) from year to year, but a large part of that may be because the rankings are done by racing year rather than by crop.  Have sires like Leroidesanimaux and Awesome Again been highly ranked for the past few years because they’re consistently getting good results, or simply because Animal Kingdom and Game On Dude have been good for several years?  Without looking at crop by crop data, it’s hard to tell.

 

While most crop-specific data is only available for purchase, there is one type that’s freely available from a variety of sources: two year old results.  I thought it would be interested to look at two consecutive years of the top two year old sires’ results and see how much consistency there was between the two crops.  By doing so, we might be able to gain some insight, although it’s unclear how much of what we’re measuring is the precocity of a sire’s offspring and how much is due to their overall quality.

 

The crops I looked at were the two year olds of 2001 and 2002.  Of the top 150 sires (by total earnings), exactly half (75) were in the top 150 both years.  I focused on those 75, and looked at the correlation on various ratings between the two years.

 

The correlations were as follows (1 indicates perfect correlation, 0 indicates no correlation, -1 would indicate perfectly negative correlation)…

Total earnings: .25

Earnings per offspring: .22

Earnings per runner: .35

Winners per offspring: .61

Winners per runner: .42

Runners per offspring: .64

So the strongest year to year correlations are between runners per offspring and winners per offspring.  That isn’t surprising, since those are the measures that are most influenced by the decisions of owners and trainers.  It’s likely that offspring of sires with reputations for siring early developers will be started earlier, while those whose sires reputations for siring late bloomers will be held back for debuts later in their two year old seasons or until they’re three years old.

 

Even the results that seem to indicate some consistency in ‘talent’ (such as earnings per runner or winners per runner) may be subject to similar biases.  Offspring who don’t begin racing until late in the year won’t have as much opportunity to include large amounts of earnings during their two year old season.  To get anything meaningful from this data, we would need (at a minimum) to know how many two year starts were made by offspring of each sire.   And (at least for the two year old crops of 2001 and 2001) that data isn’t freely available to us.

 

comments powered by Disqus

Older Comments about Year to Year Correlation of Sires' Two Year Old Offspring...

Interesting... Possibly worth prompting a database that uses this all in evaluating stallions as well as mare quality.

Related Pages


Meet Alex Zelvin

Alex is a lifelong horse racing fan and statistics nerd. He combines those interests in a passion for statistical analysis of thoroughbred pedigrees. 

Alex is currently building a model to predict the on-track success of any real or hypothetical mating. If you're interested in keeping up with his progress in-between articles here, he'll be posting frequent informal updates at his blog (thoroughmetrics.blogspot.com), where you can also sign up for his email update list. 

He can also be reached at zelvin30@hotmail.com if you're interested in consulting with him or just discussing pedigree analysis. 

Related Stories

Best of the Blogs

Top Stories