Why Roger Clemens Isn’t Barry Bonds

No matter what the outcome of the tests now being conducted on the Roger Clemens-Brian McNamee syringes and no matter what the outcome of the anti-defamation suits Clemens and McNamee have filed against each other and no matter whether the government finds grounds to indict Clemens on perjury, there’s one issue that no one is raising: Did anything Roger Clemens took or might have taken have made him a better pitcher?

A couple of ace numbers crunchers, David Ezra, author of Asterisk: Home Runs, Steroids and the Rush to Judgment, and J.C. Bradbury, author of The Baseball Economist: The Real Game Exposed, have thrown some chin music at the ongoing charges that  Clemens got any kind of statistical boost from performance-enhancing drugs.

In the mainstream media, Clemens’s guilt has long been a foregone conclusion, the case against him as airtight as the one against Barry Bonds.  But there are a few dissenters, or at least some who want more conclusive proof.

Last summer, Bob Costas outlined an argument for skepticism on Real Time with Bill Maher.  

"To much of the media, Barry Bonds and Roger Clemens make a perfect pair of bookends: one black, one white. One the greatest hitter in the modern game, the other the greatest pitcher."

But, as Costas cautioned, "Their careers are not the same.  Bonds’s numbers took off into the stratosphere after his association with BALCO. All Clemens did was maintain a career that was already on track for the Hall of Fame."

Shortly after the Maher show, I asked Costas to elaborate his point.

"I’m not saying that Clemens is guilty or not guilty of using some form of PEDs," he told me. "What I’m saying is that there’s nothing in his career numbers that would indicate any kind of artificial enhancement. I think a lot of people who assume that his later career performance was artificially enhanced haven’t done their homework." 

Bradbury and Ezra, who have done their homework, agree. "The late career spike in Clemens’ performance," says Bradbury, "just doesn’t fit McNamee’s accusations. First, the alleged use of human growth hormones took place, according to McNamee, at times when Clemens was pitching very well. Why would he have bothered?  Second, Clemens’ late success, though remarkable, is far from unprecedented.  Nolan Ryan, Curt Schilling, Randy Johnson and numerous others can be cited as examples of pitchers who excelled well past their physical prime. There’s nothing exceptional about Clemens’ last few seasons." 

According to Ezra, "Clemens was a better pitcher than Nolan Ryan, but his career swings are similar. Ryan had some relative down years at ages 29, 31, 33, and even 38, but he was durable and pitched very well from age 41-44. Clemens had down years at 33, 36, and 39.  But, like Ryan, he was very good at ages 41-43. I haven’t heard anyone throw charges of chemical enhancement at Nolan Ryan." 

The numbers support both Ezra and Bradbury. McNamee claims that Clemens’ use of HGH began in 1998, a season in which he won the American League Cy Young award with a 20-6 record, a 2.65 ERA, and led the league with 271 strikeouts.  The problem is that Clemens was even better the year before, when McNamee doesn’t claim to have supplied him with HGH.  In 1997, Clemens was 21-7 with a 2.05 ERA; he actually had more strikeouts, 292 to 271, and pitched 30 more innings, 264 to 234.7, in 1997 than in 1998.

Costas points out that discussions of Clemens’ Hall of Fame worthiness should at least begin by acknowledging that he was a legitimate candidate before McNamee became his personal trainer. "I’m really amazed," he says, "that anyone would question that Clemens was Hall of Fame worthy before 1998. He won more than 20 games four times and 18 games in three other seasons. Nearly all his best years were from 1984 through 1997." 

By any objective yardstick, Costas is right. In his first 14 seasons, Clemens led the American League in wins three times; in his last 10 seasons, after his association with McNamee, he led just once, 1998.   From 1986-1999, he led the league in ERA five times; from 1998-2005 twice. Before McNamee, Clemens was first in strikeouts four times and second four times; after McNamee, he led the AL in whiffs just once, 1998, and finished second only once, in 2002.  (Nolan Ryan led the National League in strikeouts four times in his final five seasons.) 

Measured by a favorite tool of sabermatricians, Adjusted ERA (which allows for park factors and league averages), Clemens had six of his best eight seasons before 1998; measured by another, WHIP (walks and hits per nine innings), he had all his best seasons before 1998.

Both Roger Clemens and Barry Bonds were 35 when their alleged PED use began.  Clemens continued to be a great pitcher for most of the rest of his career, but his post-35 performance isn’t in the same universe as Bonds’.  The best all-around hitting stat yet developed is Adjusted OPS, which combines on-base percentage and slugging average then adjusts them for ball parks and league averages.  By Adjusted OPS, all four of Bonds’s best seasons came from 2001 through 2004, from ages 36 through 39.  If a chemical substance affected Bonds’s performance over that span—and Ezra makes a detailed argument in Asterisk that many factors other than PEDs could have impacted Bonds’s performance—there’s no statistical evidence that Clemens got a similar boost.

Bradbury takes the argument one step further. 

"The best evidence for a power pitchers’ potency is strikeouts.  Clemens’s strikeout rate relative to the league declined as he aged. If he was getting some artificial help, wouldn’t we have expected him to have improved in this area?  The one thing that jumps out at you when you look at the numbers for Clemens’ last several seasons is not striking out batters but preventing walks. That’s the part of his game least likely to have been affected by PEDs. I think this lends support to the idea that Clemens was able to maintain effectiveness as he got older because he simply got smarter and tougher."

Benjamin Disraeli famously said that there are three kinds of lies – "lies, damned lies, and statistics." http://en.wikipedia.org/wiki/Lies,_damned_lies,_and_statistics

If the feds or Brian McNamee’s lawyers finally bring Roger Clemens down, it won’t be because of his statistics.  Why Roger Clemens Isn’t Barry Bonds