The ‘Real’ Derby Record of Today’s Top Trainers

About two weeks ago, my friend and colleague Ed DeRosa made the assertion that, contrary to popular opinion, trainer Todd Pletcher’s 1-for-31 Kentucky Derby record was perfectly acceptable.

“Admittedly, 1-for-31 isn't the most impressive stat,” DeRosa wrote on Feb. 26, “but it's also an unfair way to couch the opportunities he's had to win America's most famous race because he's only tried 12 times (but with 31 horses).”

At best (barring any dead heats for the win), Pletcher would be 12-for-31, my podcast partner argued. Furthermore, given the large fields the Run for the Roses typically attracts, is it fair to expect Pletcher to have saddled more than one Derby winner over the past 12 years?

According to Ed, no.

“Pletcher is 1-for-31 (3.2%) in an era when all starters are 13-for-246 (5.3%),” he points out. “That makes all starters not trained by Todd Pletcher 12-for-215 (5.6%). Sure, a trainer of his stature could be expected to perform a little better than all starters, but even winning the race a second time gets him to just 6.45%, and we're about to see winning it twice is no easy task let alone the charming third time that would make him a nearly 10% trainer in one of the most difficult races in the world to win.”

DeRosa then made the point mathematically, presenting an array of tables and charts that would’ve made Carl Gauss proud — and proving that Pletcher’s single win is the most likely scenario based on the final odds of his (Pletcher’s) Derby starters.

While I have no qualm with DeRosa’s math — it’s spot-on — I do take issue with his conclusion, and particularly with the way he approached the subject. Frankly, it’s the same problem I have with the winner of the National Handicapping Championship being awarded an Eclipse Award as Handicapper of the Year: In low-probability events, it is generally more illustrative to look at overall consistency rather than focus solely on winners and losers.

Don’t get me wrong: I’m not taking a shot at past and future winners of the NHC, but there is a reason that there has been no repeat champion since the tournament began in 2000 — just like the Kentucky Derby, one’s chances of winning are miniscule. Hence, judging the competence of a handicapper based on whether or not he/she won the National Handicapping Championship is silly.

Ditto winning or losing the Derby.

However, at least in regard to America’s biggest race, there is another standard. A recent study I did found a direct correlation between a horse’s final odds and its “Form Rating” (a number from 0 to 100 percent that I derived to assess a horse’s finishing position as well as its beaten lengths in a particular contest):

(Click on image to enlarge)

So, using the results obtained from the chart above, I thought it might be interesting to look at how Todd Pletcher and other top trainers performed in Louisville on the first Saturday in May. Primarily, I looked at two things:

1) The average expected Form Rating compared to the average actual Form Rating.

2) The number of times the trainer’s Derby entrants performed better than expected compared to the number of times they performed worse than expected.

The results of this study were illuminating.

(Click on image to enlarge)

Not only do we find that Todd Pletcher’s Louisville ledger is, indeed, dreadful (just 23 percent of his entrants performed better than their odds suggested), but look who’s at the top of the heap — Dale Romans, a guy who has yet to visit the Kentucky Derby winner’s circle.

(Click on image to enlarge)

Romans has saddled four Derby runners in his career and only Sharp Humor in 2006 ran worse than the tote board predicted.

Meanwhile, Pletcher can take solace from the fact that, when it comes to Derby disappointments, no one comes close to Bill Mott. In addition to being 0-for-7 in the big race, Mott’s charges have recorded just a two percent average Form Rating, despite a 21 percent expectation.