When the use of data analysis seeped into the Roger Clemens steroids saga, four Penn professors shifted their focus from the Wharton curve to the Rocket's splitter.
On Jan. 28, Hendricks Sports Management, the agency that represents the seven-time Cy Young Award-winner, released an 18,000-word statistical report aimed at disputing that Clemens had taken performance-enhancing drugs.
In Sunday's New York Times, Wharton professors Eric Bradlow, Shane Jensen, Justin Wolfers and Adi Wyner published their own findings on Clemens' career, refuting many conclusions raised in the Hendricks report. They plan to release a full 15-page article in the coming days.
"Clemens' people sought to use the statistics as evidence of his innocence, and we think they fundamentally failed to make any reasonable case there," Wolfers said.
The professors cite a selection bias in the sample of pitchers included in the report.
In explaining the precedent for Clemens' late-career success, the Hendricks study compared Clemens to three pitchers - Nolan Ryan, Curt Schilling and Randy Johnson - that remained successful into their 40s. This tactic immediately caught the professors' attention.
"You can't just cherry-pick people with late-career success and say, 'Aha! That proves it. There's other people like that,'" Bradlow said.
In their analysis, the professors compared Clemens with a wider range of pitchers that fulfilled the following criteria: at least 15 years in the big leagues since 1968, at least 10 starts each year and no fewer than 3,000 career innings pitched.
Using these parameters, they found that Clemens' career trajectory did not appear so typical.
"What emerges for pitcher after pitcher is you start off a little raw, you improve, you peak out at about [age] 31, and then you decline," Wyner said. "Clemens gets better, gets worse, then gets better again."
Representatives from Hendricks Sports Management - who did not return a phone message seeking comment - released a formal rebuttal of the professors' critiques.
"Roger Clemens is not like every other pitcher in this group," Clemens' agent Randy Hendricks said in a statement. "He is considered perhaps the best pitcher of his generation. The professors make the mistake of thinking that his career arc should look like the arc of every other pitcher in their selected group."
The professors contend that their work makes no attempt to marginalize the quality of Clemens' performance, but rather that it illustrates how anomalous his path to success was.
"We're not talking about quality; we're talking about trajectory," said Wyner, who won last year's Statistics Department fantasy baseball league.
"People do not [tend to] have his trajectory of having bad years during their prime and then great years at the end."
The Hendricks response also criticizes the professors for ignoring the changes that have taken place in baseball over the last several decades, such as "the lowering of the pitching mound [and] the tightening of the strike zone."
According to Wolfers, such an argument is ludicrous.
"Things like the tightening of the strike zone actually make it even more unlikely that Clemems would have improved as he got older," said the professor, who made the sports pages last year for his study on potential racial biases in NBA referees. "In fact, the criticism goes exactly the opposite direction from what they wanted to do."
Further, the professors argue that there is no statistical basis for Hendricks' conclusion about Clemens' efforts to "adjust his pitching style." One of the chief points raised in the Hendricks report was that the development of Clemens' signature split-fingered fastball - and not performance-enhancing drugs - was responsible for the pitcher's longevity.
"There's absolutely no data in the Clemens report about any kind of evolution of pitch selection," Jensen said. "For something that's supposed to be a statistical analysis that makes this claim, it would have been nice to see some statistics to back that up."
Despite their problems with the presentation of facts in the Hendricks report, the professors make no insinuations about Clemens and potential steroid use.
"I'm a huge Roger Clemens fan," Bradlow said. "We have no ax to grind. Our New York Times article is merely a factual statement about what the data says."
With Clemens' long-awaited testimony before Congress looming tomorrow morning, that "factual statement" has landed the professors at the center of a media circus.
Jensen, for example, appeared on ESPN's First Take and on the Comcast Network yesterday.
"It's been kind of a whirlwind relative to what I'm used to in my career," said Jensen. "But I think we're at the four-minute, 50-second mark of our five minutes of fame."
