Recently Roger Clemens' agents from Hendricks Sports Management published a report in support of Clemens' claim that his late career surge wasn't unusual.
The report is loaded with data on Clemens' ERA and other pitching stats and tells a detailed story of his pitching career. It also compares Clemens' performance to other pitching greats like Curt Schilling and Randy Johnson. It's a fun read.
One suggestion to the authors: if you're going to publish a report loaded with data on why your client's performance is not statistically significantly different from the average for other great pitchers at the same point in their careers, you might want to get some professional statisticians to do the analysis.
Here are some links to the bios of the authors of the report, and the degrees their bios say they earned:
Randal Hendricks: Finance, Law degrees
Steve Mann: International Relations, Educational Philosophy degrees
Brett Larson Hendricks: History, Law degrees
These are well educated guys, and they've done very well for themselves. But they don't appear to be mathematicians. Uh oh.
Several Wharton Business School Professors went ahead and checked the methods used in the report, and published their response in the New York Times today. One of the authors is Eric Bradlow, a professor I knew while at Wharton.
Take a look at his bio here. Notice the first thing under his "research areas": Bayesian Computation. What the hec does "Bayesian Computation" mean? It means GAME OVER for Clemens' non-math-genius agents who published that report.
Some takeaways from the response:
+ The Clemens report only compared his performance against pitchers who were successful in the second half of their careers, versus all pitchers who had a successful first half. In other words, the Clemens Report better supports the hypothesis that Clemens' performance in the 2nd half of his career wasn't much different than Schilling's or Johnson's.
+ The authors did some analysis using 31 pitchers who pitched at least 10 games a season for 15 or more seasons since 1968. When comparing to this group, Clemens really stands out. Most of these pitchers in the sample group peaked around age 30 and then went downhill from there.
All I can say is: OWNED.
