More On Roger Clemens

Last week, Justin Wolfers offered an insightful analysis of Roger Clemens‘s career statistics and what those statistics imply about the likelihood that Clemens used steroids.

The latest contribution to this debate is by sabermetric legend Nate Silver of Baseball Prospectus. Using only data through 1997, Nate generates a projection of what Clemens‘s stats should have looked like from 1998-2001, the years in which the Mitchell report alleges that Clemens used performance enhancers. Silver concludes that there is nothing particularly unusual about the pitcher’s performance during this four-year period.

Does this conclusion vindicate Clemens? Not by a long shot. While statistical evidence can sometimes provide convincing evidence that something really out of the ordinary has happened (like the sumo wrestling cheating that was documented in Freakonomics), it is far from clear how to interpret the findings when things look normal. Indeed, most of the studies that have examined data on steroid use in baseball have found little evidence that steroids actually enhance performance. Stephen Stigler, a statistics professor here at the University of Chicago, came to that conclusion. So does Nate Silver in a prior analysis. The only exception I know is Ray Fair‘s study on age effects in baseball, in which he shows that almost all of the batters in his sample who far exceed his predictions late in their careers played during the 1990s.

Data are great, but in this instance analyzing dirty syringes or subjecting old blood samples to new doping tests are both more likely to provide definitive answers.

(Hat tip to Ken Kovash)

Joe T

Scott is is obviously "not in perfect mental balance" (which I think is this week's politically correct euphemism for "a stark raving lunatic"), but for the record, it should be noted that "roid rage" is a media made up phenomenon with no basis in science.

Not coincidentally, the same is true for road rage, the fake pop psych term from which "roid rage" drew its name.

Donald A. Coffin

Let's see. Wolfers' analysis is "insightful," while Silver's analysis doesn't provide a useful contribution to the debate?

Wolfers' analysis seems to support the contention that Clemens used PEDs, while Silver's would raise doubts.

So...what is your prior here? Seems to be that Clemens used PEDs.

Kevin L.

If the data revealed shows that steroids aren't enhancing performance, then what are the players' incentives for using it? If you showed a player that he was playing about the same before and after using the steroids, then would they still use it?

Now I know we're trying to figure out if these players used steroids in the first place. But let's say we live in a world where steroids were natural and legal. And let's say that the accused all did use steroids. So what this post is saying is that the substance really did nothing for the players. I'm not refuting Levitt's claim that lack of change in the data should vindicate the players. I'm asking why isn't the performance data showing anything. If people can show that steroids don't help, then we would stripping the perceived effectiveness and need to take performance-enhancing drugs that don't really enhance.

Kyle S

Ray Fair's initial study had a number of problems with it as chronicled by Guy on the thread below:

Perhaps he's fixed it by now? I don't have time to check.


Apropos of Stigler's study: I recall reading about the non-influence of steroids on performance back in the McGwire controversy. Seems that players take this stuff because they believe it enhances their performance; i.e., taking steroids is a superstition, like not stepping on the foul line when running to or from the dugout. If a player believes steroids are helping, are they?


Steroids and HGH do more than just increase existing performance, they also help reduce recovery time. As a former college athlete in the 90's, I also know there was a belief that steroids, and especially HGH, would help reduce nagging muscular and tendon injuries, such as torn rotator cuffs. I don't know if there is science to back that up, but it was perceived. I personally know of one guy who took HGH just to help him recover from some minor nagging injuries during the season - so he took them just to get himself back to equal with past performances.


Clemens won't be vindicated until he apologizes and the fans forgive him- the irony of clinging to a lie is that it alienates him from his own supporters, who are forced to choose between admiration of the athlete, or indignation at the perjurous pride

Tom W

Scott S; What a novel and sophisticated approach...Would that everything in life was so easy. Apply the simple litmus test of support for George Bush to determine who is a jerk and who is not.

ML Harris

Interesting: The frequently jaundiced eye of economics sees statistics as able to prove cheating but not to disprove it.

I don't mean to take the science down a peg, but last week, we were convicting Clemens because one study (flawed due to the statistics used). Today, we are leaving him convicted, despite a reasonable study suggesting otherwise.

I don't really care if Clemens used. Or Bonds for that matter. Detestable though they may be, it's in the way of my enjoyment of the strategy and majesty of the game.


It would be easy if people just told the truth.


"Does the intent to get ahead illegally without actually benefiting count as cheating? "

-Ask Pete Rose what he thinks...not only didn't 'cheat' persay but also (as far as I've read) only bet FOR his team to win. Nonetheless, deemed so unsavory he's been banished for life.


ML, studies rarely disprove anything because it's very hard to prove a negative.

And if studies show that using steroids doesn't tend to improve performance, then this analysis is even less likely to disprove that Clemens used PEDs.


Brian: if i can bet on myself then i can still perform poorly in order to improve my return from betting. i can sabotage the teams performance for a number of games when i'm not betting and then when i want to bet i stop sabotaging and receive better odds than i would have normally received.


Joe (#13) ... Anybody who doesn't believe in road rage has never commuted on the DC Beltway.

Roger Clemens' performance had started to dip a little when he moved to Toronto, at about the time he instituted a workout regimen that let him stay effective (unless anybody can point me to some old, fat, out-of-shape power pitchers) without becoming a soft-toss artist like Mussina, Perry, Niekro or Blyleven, considered by statistics as "comparable".
As Chris (#9) noted, steroids are supposed to increase the ability to recover from workouts, fatigue, injury, etc. So it would be hard to imagine a statistical test that could determine whether any pitcher used them. I mean, the difference between these guys and the less comparable ones who flamed out at 34 is that these guys didn't break down ...
It would be interesting to see a "most comparable" list of great young power pitchers who ate themselves into injury and retirement by their early 30s.



@3: Does the intent to get ahead illegally without actually benefiting count as cheating?

This is an extremely interesting philosophical question. For example, intent to murder is not punished to the same degree as murder. Suppose I am distracted and momentarily lose control of my car. I drive off the side of the road. If some one happens to be there, I may be convicted of vehicular manslaughter. If no one is there, I pay a tow charge. My intent was the same. It was only the circumstance that was different.

I study cheating on tests. Some people feel that cheating without passing is really not cheating. Others acknowledge that cheating without passing is cheating but it shouldn't be punished. And a third group feel that cheating is cheating regardless of whether the cheating was effective or not.

The question above implies that if the steroids didn't provide an advantage, then it was just stupid cheating and it didn't really matter because no one was affected. Of course, everyone is affected when people cheat. We no longer trust the cheater. If they will lie, cheat and steal about one area of their life, we are inclined to believe they will do so in another area of their life.



Instead of trying to measure E.R.A. which does have an influence on the team's defense and situational pitching why don't we measure average length of starts, number of pitches thrown, average velocity of fastball?

For batters I think you would have to measure recovery time from injuries, average bat speed from before and after supposed steroid use.


There is no way statistics can prove or disprove steroid/hgh usage. You are by definition dealing with people on the extreme end of the bell curve of athletic ability. So even narrowing it to those few that may be a close fit during the early part of thier careers, you have an extremely small sample size and can expect huge variations based on longevity factors that might be unrelated to the talents that got them there in the first place. It provides an interesting framework, but can't be taken as proof one way or the other.


If steroids doesn't affect performance, should we care? Does the intent to get ahead illegally without actually benefiting count as cheating?

Scott Supak

When Clemons threw that broken bat fragment at Piazza, it seemed like a lot of rage for a guy who's supposed to be a professional.

Then I found out he supported the lying cheater George W Bush, and I thought, well, that tells me something about him, doesn't it? If it's OK for Bush to lie and cheat, then why not Clemens? Why not anybody?

Roid's or not, the guy is obviously a jerk.

If it turns out he lied to congress, I'll expect the same yawn that his President got when he lied to congress.

But if you or I did it, boy howdy!


There may be no statistical evidence of enhanced performance, per se, but that doesn't necessarily disprove anything.

Maybe athletes take it just to stay even from the prior year's performance, i.e., you won't see "enhanced" year-over-year statistics, but taking the steroids help the aging athlete remain at his previous levels and not show a decline.