Defending Roger Clemens

Today, Allen Barra quotes me extensively in a NY Observer article defending Roger Clemens.

“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.”

If you would like to see my analysis of Roger Clemens’s career, you can read my initial post on the subject here. And, this post contains links to several statistical analyses of Clemens’s career. In summary, my view is this: if Clemens used performance-enhancing drugs, they didn’t work. Contrary to popular opinion, the statistical record does not support Brian McNamee’s allegations.

16 Responses “Defending Roger Clemens”

  1. Josh Lipman says:

    What about drugs like HGH that allow faster recovery? I know there’s much evidence contradicting the efficacy of HGH, but I’m sure there are similar, illegal drugs whose effects can’t be captured with performance-related statistics.

    Also, is it possible that though his strikeouts decreased, they were decreasing at a slower rate than would have occured naturally? And, even so, is it possible that he started pitching to contact more, as he became older and presumably smarter?

    I just don’t think it’s possible to accurately assess steroid use – or lack there of – through strikeouts only.

  2. Mitch says:

    Steroid use could also prevent a decline, in which case he might decline relative to the league but outperform relative to an age adjusted baseline.  And, for the record, I don’t care one bit about steroids, I just don’t think your explanation would account for this possibility.

  3. Millsy says:

    My theory is it doesn’t really matter.  Players have long been taking ‘PEDs’, whether it’s caffeine, novicane, cocaine, steroids, etc.  The truth is this era isn’t any less ‘valuable’ than any other.  There is never going to be any way to parse out the effects of steroids anyway.  Nutrition, weight lifting, etc. are also increased dramatically from any other era.  And the kicker is: players are doing it on both sides of the ball.  The relative competition is the same, but the quality of the game has increased dramatically.  Tell me you didn’t like Josh Hamilton hitting balls off the back wall of Yankee Stadium.

    And on another note, why hasn’t anyone chastised football?  Has anyone actually looked at those guys?  Maybe it’s my HDTV, but since when can we see veins on a natural person from 50 yards away?  I’ve seen some amazing athletes in my endeavors in sports (none of which looked anywhere close to these guys), and there seems to be a huge jump in what they actually look like when they get to that level.  Something changes.  I say take the eye off of baseball for a while.

    And what’s a PED exactly?  If A-Rod has a 5 hour energy before the game is that a PED?  I bet excessive use of that causes more heart attacks than steroids (ephedra anyone?).  It’s just a game that goes back and forth.  The fact that Congress is wasting so much money on this issue makes me angry, especially in this climate when they should be discussing NOT giving away $2 trillion to bad businessmen.

  4. Dan says:

    In my mind – steriods biggest impact are the aid to quick recovery from injury especially for older pitchers (who are particularly injury prone).  The fact that Cleamens was in decent health into his 40s gives you evidence to suspect him (but of course not to convict him). 

  5. werr says:

    It’s quite possible that Clemens juiced long before he met Mcnamee.

  6. JC says:

    From my original post, linked in the above article.

    There is also the possibility that Clemens used steroids throughout his career, and that other pitchers having late-career success have all been aided by performance-enhancing drugs. I find both to be unlikely, especially the former. If Clemens continued employing McNamee several years beyond the last time the alleged steroid injections occurred, why would he go somewhere else for his steroids? Or, why would McNamee lie to say that Clemens was clean during this period?

  7. Stephen says:

    I don’t know anything about McNamee so I won’t make any arguments about lying or dates.

    But I am confused about the answer to Dan (#4)’s argument that perhaps steroids prevented a decline.  The fitted value plots are visually interesting but they don’t shed any light on whether or not drugs prevented a decline relative to some benchmark.

    Very, very crudely, it’s like you want a setup where you estimate a model [whatever statistic] as a function of age and see if a Clemens dummy is significant.

    Specifically regarding the quote from JC’s original post:
    “Other pitchers having late career success used, too”
    – You haven’t shown us this late career success other than to name drop.  The names are persuasive, but not a very large N, and I’d imagine there are a lot more late-career flameouts than Nolan Ryans.
    – My suggested “model” implicitly assumes that a large number, those making up the baseline, didn’t.  Something to keep in mind, an assumption which may or may not be true.
    – Another shortcoming of my proposal is that it doesn’t provide any causality.  Specifically, it doesn’t rule out that Clemens may just be a total man-freak of a pitching machine, which is entirely possible.

    If we don’t find support for the hypothesis that Clemens’ decline was slower than comparable pitchers, then JC’s argument is all the more persuasive.

    If we do find support, then we’d have to think about this more.

  8. Marc Schneider says:

    I’m not sure what difference it makes whether his performance improved or not.  He may well have taken steroids and they did not help.  The issue at present is whether he lied about taking them.  JC’s point goes to whether Clemens’s credentials are valid in terms of the Hall of Fame but not necessarily as to whether he is lying or not.  Same with Bonds. 

  9. Fletch says:


    I find your argument compelling as far as it goes, especially in relation to those who argue that the statistical variability favors the prosecution.  I am particularly in agreement that an assessment based upon DIPS, and strike-outs in partcular, rather than WHIP (defense-correlated) and/or references to general health (which varies far beyond any useful analytical toll and is thus useless as evidence).  There is merit to the argument that a league-related adjustment either to the mean or comparison sample might be justified based on the talent faced but I doubt that would significantly alter the analysis.

    It sems – and I’m new here just going through the whole thing for the first time – that most of the critics are either attempting to read more into the analysis than can be found (how can you say unequivocally that Clemens is innocent?!?!) or are going beyond the analysis to make subjective assessments of the credibilty of the parties involved.  I was particularly amused by those arguing that McNamee could not possibly risk jail to throw Clemens to the lions for his own benefit.  Such risk is obviously non-existent as he knew that he had injected other substances (B-12 is frequently self-administered, not just by doctors) and it is a practical impossibility to prove a negative.  Likewise, the refusal to take a (notoriously unreliable) lie detector test is evidence of absolutely nothing.

    I commend the effort at the very least.  Having tried to explain the lagging nature of unemployment (or, worse yet, Fed policy) to people with only a popular unexamined understanding of the issues (who have largely made up ther minds already), I know what you face.

    Personally, I agree with you that the statistics do not support an assertion that Clemens’ performance was in any measurable way affected, but that his subsequent actions/statements have, at least, been troubling.  Thus suspicion is, perhaps, warranted, nothing conclusive (or damning) has yet seen the light of day.

  10. Rodrigo Coutinho says:

    Mr. Bradbury,

    First of all, sorry about my english. Im from Brazil. I usually read your blog. I like your points of view a lot, even when I dont agree with them. Talking about drugs who enhace the players perfomance, and their consequences, for example, use are preventing players to be elected to HOF (McGwire…), in “defense” of players you keep saying that taking that drugs wasnt illegal at the time.

    I did read today at SI Site this statement:

    “Though MLB’s drug policy has expressly prohibited the use of steroids without a valid prescription since 1991, there were no penalties for a positive test in 2003. ”

    If this is true it was illegal to take that drugs. The unique thing is there wasnt consequences for doing something wrong, but the procedure was forbidden.

    So, this would change the things a little bit, right?

    I think this debate will be hot now with A-Rod news.



    PS: I bought your book last time I went to USA. I love it.

  11. JC says:

    Thanks Rodrigo,

    That statement refers to a memo by Fay Vincent. It was not binding, because it was not approved through the collective bargaining process. The Commissioner did not, and still does not, have the power to unilaterally implement such a rule. It would be like claiming a bill that passed the Senate but not by the House or the President was a law.

  12. werr says:

    Responding to post #6.  Don’t the effects of steroids last for years?  Maybe Clemens still benefited from the gains he made from steroids many years later,  if he used them of-course.

  13. JC says:

    The effects evaporate rather quickly after a player stops using, according to my exercise physiology colleagues.

  14. werr says:

    Could you keep the gains if you used other supplements like Creatine?

  15. JC says:

    I doubt it considering that creatine really doesn’t help much, if at all.

  16. Rodrigo Coutinho says:

    Creatine isnt considered a PED.