The Clemens Report

Proving that his competitive drive is stronger than ever, Roger Clemens has released his own report. At 49 pages, it's a bit short of George Mitchell's, but it promises to be just as interesting a read. In it, he makes the case that his trendlines are not as anomalous as his critics (including myself) have argued.

Update: My take on the report after the jump.

My first beef is the attempt to compare Clemens' "ERA margin" (essentially, ERA+) to the careers of Randy Johnson, Curt Schilling and Nolan Ryan. IIt argues that all four saw great ERA margins in their late 30s.

To the report's credit, it shows the line graph of each player's career -- but that ultimately disproves the notion. It's pretty clear that Johnson and Schilling both had pretty obvious, though gradual, declines after a single peak, but Clemens features three separate peaks, the latter two of which are higher than his early career peak (which you'd assume would be highest). Ryan's is the most similar to Clemens', except that his late--career peaks match his early-career peaks -- they don't exceed them by a wide margin, like Clemens' does.

Later, it propogates the "twilight of his career" myth, trying to rebut the notion that Clemens was in his twilight from 1994-96 when Dan Duquette never said any such thing, talking instead about Clemens' likely future at the time: "We had hoped to keep him in Boston during the twilight of his career."

It goes on to attempt to show that Clemens was comparable to eight other HOF pitchers just on his Boston numbers alone, but the list is populated mostly by deadballers and borderline selections, and while Sandy Koufax is a big name, he's in the Hall for putting up single-season and postseason numbers Clemens never approached in Boston.

One of the better charts is a comparison of Clemens, Johnson and Schilling's K rates by age. The report notes that each pitcher recorded his highest rates in his late 30s, and that Johnson's 10.6 K/9 at age 40 was higher than any K/9 rate Clemens put up. Which is kind of the problem. Johnson was clearly a much bigger strikeout pitcher than the other two. For him, 10.6 isn't even one of his 10 best seasons. But the 9.2 Clemens posted at age 41 was his sixth-best -- just outside the somewhat arbitrary highlighted top-five-seasons zone chosen by the report's authors. Also unmentioned in the report is that Clemens' K rate was declining in his late 20s before spiking slightly in 1994 then spiking again tremendously in 1996. (The report does attribute Clemens' poor performance in the mid-90s to injury, but this is an assertion easily made and more difficult to prove). Johnson and Schilling both saw steady increases into their early 30s, when they began posting monstrous K rates.

The argument really seems to come down to whether Clemens is a freak like Nolan Ryan was. Ryan posted his three best K rates at 40+ after nearly a decade of lesser rates. "It should be noted that Roger Clemens has stated for many years that his idol was Nolan Ryan," and that he emulated Ryan's workout style and mechanics, the report says.

It goes on to compare Clemes' IP/G and Pit/G totals to Johnson's and Schilling's. It shows that all three were fairly consistent and peaked in their late 30s. I'm not sure how relevant this is. The report says it shows stamina, but innings and pitches tend these days to be more closely related to performance, not stamina. And we already knew these three pitchers performed well in their late 30s. That's not the question. Though to his credit, Clemens clearly did have the stamina to pitch for a long time, and that never declined until very recently.

The report wraps with a list of Hall of Fame pitchers who pitched into their 40s, which is frankly unhelpful and misleading. Clemens pitched (assuming he's done) until he was 44. Yet the list includes seven pitchers who retired at 40, two knuckleballers, seven deadballers (by my count) -- never mind the pitchers like Lefty Grove, who retired at age 41 but had zero effective seasons after age 39.

In all, it's a valiant effort, but it's not going to change anyone's mind one way or the other. It doesn't answer my questions about Clemens' trajectory, and essentially hinges the statistical case on Clemens being the new Nolan Ryan, whose career was unique in every sense of the word. For those who believe Clemens is telling the truth, there's plenty of evidence here that presented the right way can bolster that supposition. It just doesn't hold water for me.

Comments

[Lar]

I'm sure roids help, but obviously there's not enough control to see. I wonder about the placebo effect..

[walein]

Saber guys aren't going to be happy about the heavy use of ERA as this report's important tool to distinguish Clemens as a pitcher.
It's also hilarious that they've used this report to show how Curt Schilling and Randy Johnson may be more likely candidates for "performance enhancement" through their "ERA Margins."

[Paul SF]

Updated with my thoughts up top.

[Nick-YF]

"and while Sandy Koufax is a big name, he's in the Hall for putting up single-season and postseason numbers Clemens never approached in Boston."

I can't speak for the post-season numbers, but doesn't Clemens peak with the Sox (1990-1992) approach Koufax's 1964-1966? And doesn't his his three year stretch from 86-88 best any 3-year combination in Koufax's career pre-1964?

Overall, as Paul writes, this report probably isn't going to persuade those who think Clemens is guilty that he 1. didn't use and that 2. PEDs did not improve his performance. Still, it's part of pretty aggressive public relations strategy, and I don't think it's harming Clemens in any way. In these times, silence is equated to tacit admission. If you're innocent (and I'm not saying Clemens is one thing either way), you might as well be loud about it.

[Paul SF]

My big issue with the report is it's stretching the bounds of analysis to claim what Clemens' agents want to claim -- that his career is similar to that of contemporary high-K pitchers.

That's just not true. Clemens has had a unique career, and it would behoove his defenders to admit that and come up with one of the several plausible explanations for that uniqueness. The one that makes the most sense is that injuries hurt his abilities in the mid 1990s, and his recovery coincided with his mastery of the splitter to return him to dominance in the late 1990s.

And although the report is claiming that's what happened, it's alsot rying to make Clemens' overall trajectory seem relatively ordinary, and it's just plain not.

[Paul SF]

Also, a quick and dirty sketch of Koufax v. Clemens:

Best three year period, by ERA+

Clemens:
1989 -- 132
1990 -- 213
1991 -- 164
1992 -- 175

Koufax:
1963 -- 159
1964 -- 187
1965 -- 160
1966 -- 190

So they are somewhat comparable (particularly when you look at three-year peak), and Clemens has the top season, so I overstated my case. But when you add the fourth year, I think my general point stands.

They both pale to Pedro though:

1997 -- 219
1998 -- 163
1999 -- 243
2000 -- 291
2001 -- 189
2002 -- 202
2003 -- 210

[YF]

"Clemens has had a unique career, and it would behoove his defenders to admit that and come up with one of the several plausible explanations for that uniqueness."

As Walker suggests in the thread below, perhaps there is no need for him to explain this variance, that the spikes in his performance are simply natural statistical outcomes for someone with his extraordinary innate skillset.

Again, I hate to speculate about this, as:

(a) there is no actual proof that he took any drug
(b) there is no study as to the effectiveness of any drug he might have taken

[Paul SF]

Well, they could say that, it seems like. All this business about Johnson's career and Schilling's career is just noise. If the Hendricks brothers wanted to release a report focusing solely on how Clemens' slump in the mid 1990s wasn't as bad as perceived, and then went into more detail on Clemens' pitches before and after that gap, and talked about statistical probability of having a moster season like he had in 1997 given the results from 1993-96, that would be a study worth seeing.

If I remember correctly, there are some pitchers in history that have career trajectories more similar to Clemens than Johnson and Schilling. I guess I'm wishing that the Clemens Report, like the Mitchell Report, had a little more substance.

[Jonathan]

Isn't the whole point of the investigation based on the statement by his trainer that he administered steroids to to Clemens? I guess I falsely assumed that a report by the Clemens camp would rebut that accusation by finding evidence that the trainer was lying, or find means to prove Clemens was innocent. Not provide a statistical package as a defense. IMO, that document is simply 49 pages of spin. Was Clemens great, yes no one can argue with that. The real question is did he take PED's. From what I understand, the 49 page document doesn't prove that at all.

In reality, no one can predict what he would or would not have done while taking PED's.

Very disappointing IMO. Sad.