Baseball Crank
Covering the Front and Back Pages of the Newspaper
January 4, 2005
BASEBALL: Yes, Steroids Help

Unless you take the strong libertarian position - that there should be no restrictions on what ballplayers can ingest regardless of the impact on themselves or the game - the debate about what to do about steroids in baseball really revolves around three questions:

1. Does taking steroids help make you a better baseball player? (If not, there's no point in banning them).

2. Is taking steroids harmful to your health? (Again, if not, there's no reason to ban them)

3. Is there a feasible way to test for steroid use or otherwise enforce a ban?

I recognize that there are serious people who disagree about the second and third questions. But I submit that, if you think about it honestly, what we do know about the first point is quite clear: steroids* can and do help performance in baseball, and specifically help in hitting for power.

* - I refer here colloquially to "steroids" to include other hormone-altering performance-enhancers like human growth hormone. As often happens in debates about drugs, precise definition of the substances involved is itself a whole sub-field of debate.

The Available Types of Evidence

Part of the confusion over the link between steroids and performance derives from the different types of evidence we use to answer these types of questions. To illustrate, let's compare this question to one with a settled answer: whether throwing the ball faster will help a pitcher strike out more batters.

Direct Evidence

One sometimes hears the argument made that we can't and don't have direct evidence of how steroids help performance. This is true enough, as far as it goes. For example, we can show directly how velocity helps a pitcher get strikeouts: you can measure batters' reaction times and show how increasing velocity makes it harder to make contact. Or, you can simply watch a guy who throws 95+ blow pitches even past guys who are looking for them. That kind of "see the causation with your own eyes" evidence doesn't exist for steroids and performance in baseball.

Statistical Proof

Where direct evidence of causation isn't available, of course, statistical proof of correlation can be good enough. A classic example of this from the intersection of law and medicine is the fact that we still don't have direct evidence that smoking cigarettes causes lung cancer (i.e., scientists can't show how it happens), but the statistical evidence shows a fairly overwhelming connection between smoking and increased likelihood of getting lung cancer.

Statistical proofs of correlation are pervasive in baseball - to use our example above, it would be easy to do a study showing that pitchers who regularly throw above 95 mph get a lot more strikeouts, and are much more likely to generate large numbers of strikeouts, than pitchers who rarely or never crack 90+ mph. That correlation is so powerful that it will show up in almost any study.

Other correlations are trickier, which is why a reliable study has to use a large enough sample size to be able to generalize, and has to ensure that truly comparable players are being compared, so that different outcomes can't be explained away by some other factor.

Here, there are two problems with studying steroid use. One is finding large and otherwise truly comparable sets of players (comparing the same player before and after isn't useful because of the interfering factor of age, which ordinarily is, of course, very powerfully correlated with declining performance after about age 28 or so). But the bigger problem is that steroid use, by virtue of being illegal, is done in secret; we have so little reliable information about who uses what, when and in what amounts that for the foreseeable future, it will be impossible to do statistical comparisons with any degree of confidence.

Circumstantial/Inferential Evidence

The fallacy in many arguments over steroids in baseball is to note the lack of direct or reliable statistical evidence and declare the question unresolved. But this is not consistent with how human beings make decisions in everyday life, in law, medicine, politics or in baseball. When the best forms of evidence are unavailable, we look at what remains: at circumstantial evidence, and logical connections to be drawn therefrom. For example, even if we couldn't see fast pitches going by hitters and read the evidence of the same in box scores, what we do know about hitting a baseball - you have to time your swing to make contact - is itself strongly suggestive of the fact that a faster pitch will be harder to hit.

I would submit that that evidence is more than sufficient to persuade us that steroids help performance in baseball. Let's once again break this down to a few questions:

A. Do Steroids Help Build Strength?

This much is not seriously disputed, which is one reason why steroids are banned in the NFL and the Olympics, where physical strength and speed can be shown to connect directly to performance. There are certainly debates about precisely how and to what extent steroids help, but few serious people would debate that taking them helps build stronger muscles.

B. Does Strength Help In Hitting A Baseball?

This is really the crux of the argument. It is often said that you can't take a drug to help you hit a curveball, which is true but totally beside the point. The issue isn't whether steroids will help you or me become a major league ballplayer; the issue is whether guys with the pre-existing skills to play professional baseball will have those skills enhanced. To deny that, among other things, you have to argue that strength has no impact on the ability to hit for power. Of course, this is ridiculous. Since the introduction of the home run as a regular part of the game in the 1920s, it has always been the case that big, strong guys with powerful chests and arms have tended to be home run hitters, and skinny little guys have not. To deny that steroids have an impact on hitting for power in particular, you have to look at all the home runs hit by the Gehrigs and Foxxes and Mantles and Kluzewskis and Killebrews and all the singles hit by the Willie McGees and Vince Colemans and Nellie Foxes of the world, and argue that it is just a coincidence that physical strength has always been so strongly correlated with home run power. You have to not only look at Bonds and Giambi and all the other guys who have been placed under one sort of cloud or other and say that whatever they took or were given didn't matter; you actually have to say that all the muscle Barry Bonds has added has had nothing to do with his power surge, that Jason Giambi's increased power production as he gained muscle was just a coincidence. Sorry, I'm not buying that.

Basic physics: force equals mass times velocity acceleration. The force you hit a baseball with is affected by the weight and speed of the bat. Stronger players can generate greater bat speed, or generate the same bat speed with a heavier bat. Yes, bat speed is a variable affected by other factors - the arc of your swing, reflexes/reaction times . . . and yes, it's true that muscle mass sometimes gets in the way of greater bat speed. But again: if strength has nothing to do with power, why have stronger players always, as a class, hit for more power?

C. Do Steroids Help In Conditioning?

Strength is the core of the debate. But correct me if I'm wrong here - I believe most of the analyses I've seen have similarly shown that steroids can assist more broadly in conditioning - beyond pure muscle mass - by assisting in the ability to train at greater length without injury, at least in the short run.

D. Does Superior Conditioning Help In Baseball?

The question, again, essentially answers itself, and doubly so for aging players seeking to stave off declining bat speed (or declining velocity, for pitchers, but pitchers and steroids are another day's debate). Honus Wagner lifted weights; Ty Cobb was a conditioning fanatic. It could be a coincidence that they lasted into their 40s in a day when few others did.

The Bonds Issue

I would stress, again, that I don't have anything but the sketchy information in the public record on what Barry Bonds took and when, and how it helped him. And it's true: Bonds' late career surge has had other causes, from better bats to a greater uppercut in his swing. But I've been disappointed at some of the efforts from otherwise reasonable people to obscure the fact that Bonds' increased strength has had an impact on his unprecedented late-30s power surge.

I meant to get to this when it ran in mid-December: the New York Times editorial by Will Carroll of the Baseball Prospectus (discussed here on his blog). I like and respect Carroll from his work at BP, but the Times piece has some serious issues. One is the point I make above: Carroll essentially implies that he is agnostic on whether strength helps with power hitting, contrary to 85 years' experience:

[W]e have little or no idea what these drugs accomplish. Do stronger players hit the ball farther, swing the bat quicker or throw the ball harder? Does using steroids reduce fatigue so that they can do any of those things more effectively than "clean" players?

While there is no doubt that these chemicals are effective at their stated goal, albeit with significant complications, the question of how their effects manifest themselves in a baseball game has not been answered. There are no credible studies that connect drug use to improved performance, nor any that determine what cost these athletes may be paying.

Much more problematically, Carroll uses some seriously misguided examples to imply to the Times' readers that Bonds' power surge is not so unprecedented:

It is true that Bonds's performance over what many would expect to be the twilight of his career has been incredible. Instead of a slow decline as he approached 40, Bonds has done what can only now be described as superhuman. . . . The raw numbers, however, only reflect his increased home-run production; they do not say whether he hits more homers that fly significantly farther.

What of this late-career surge? Certainly we can point to that with an accusing finger, sure that Bonds's numbers in the record books have been written with some "cream" or "clear" substance. It's much easier to point than to find facts.

According to Clay Davenport, a researcher at Baseball Prospectus, Hank Aaron's best year for home runs - when he had the most homers per at bat - was 1973, when he was 39. His second best was in 1971, at age 37. Willie Stargell had his best seasons after age 37. Carlton Fisk put his best rate in the books when he was 40. Even Ty Cobb had his best home run rate at age 38, though the end of the dead-ball era helped that. It is not uncommon, according to Mr. Davenport, for a slugger to change his mechanics as he ages, swinging for the fences as his ability to run the bases declines.

These are terribly bad examples. First of all, Aaron in 1973, Stargell in 1978 and 1979 and Fisk in 1988 all had one thing in common: none of them were full-time, 500+ at bat players any longer, as they'd been in their primes. It's a lot easier for an older player to improve his production if he has a third to half of the season to rest as opposed to the years when he was playing every day, a fact that has absolutely zero to do with Barry Bonds.

Let's take Stargell first, as he's the most egregious example. Willie Stargell's career best slugging percentages, both absolutely and relative to the league, came at the ages of 26, 31, and 33, well within the normal range. Stargell's home run rate improved slightly in 1978-79, at age 38 and 39, but his doubles - also a key power stat - dropped off sharply from 43 in 1973 to 18 and 19 in 1978 and 1979. Was he really hitting for more power? Also, Stargell had another thing going for him: while he wasn't, strictly speaking, platooned (his backup, John Milner, was also lefthanded), the decline in his playing time allowed him to see a much more favorable mix of pitchers: Stargell had 30.5% of his at bats against lefties in 1978 and 30.7% in 1979, as opposed to 39.5% in 1971 and 33.1% in 1973. For a guy with Stargell's big platoon splits, that's a significant advantage.

Then there's Aaron. If you know the game's history, you already know that Aaron's late-career power surge was an illusion created by the improved offensive conditions of the 1970s as opposed to the 1960s, combined with his move in 1966 into homer-friendly Fulton County Stadium and out of pitcher-friendly Milwaukee County. Aaron hit 52 homers on the road and 37 at home in 1962-63; in 1971 and 1973, those figures were more than reversed to 55 at home and 32 on the road. But it doesn't stop there; with just 392 at bats in 1973 at age 39, the right-handed Aaron saw 44.4% of his at bats against left-handed pitching, up from 30.9% in 1971 and 26.5% as a full-time player in 1969.

Then there's Fisk, whose "best" home run season was 253 at bats in 1988. Do I really need to explain why a catcher might hit better playing half the time? And yes, the right-handed Fisk faced lefties 36.5% of the time in 1988, compared to 22.9% in his actual best season, 1977.

(Ty Cobb, whose career high in home runs was 12 but whose career high in slugging average was at age 24, is not even worthy of discussing at length).

None of these guys - indeed, no other player in baseball history - compares remotely to what Barry Bonds has done, and it does no service to the debate to pretend otherwise. Prior to 2000, Bonds was 34 years old and had a career slugging percentage of .559, with his two best slugging percentages (.677 and .647) coming at age 28 and 29. Since then, he has slugged .781, a 40% improvement on his career average and a 15% improvement over a five-year stretch compared to his career best season. Neither Carroll nor Davenport could find an example anywhere, certainly not outside of guys who straddled the arrival of the lively ball in the 1920s, of an established player who had anything like a 40% improvement in his power numbers from age 35 to 39. (Bonds has also batted .358 over the past three years, compared to batting above .320 just once through age 35, also nothing like a normal aging pattern).

Carroll's argument would have been better served by recognizing the fact that what Bonds has done is totally unprecedented and clearly not unrelated to his dramatic improvement in physical strength in his late 30s. Pretending otherwise does no one any good.

Posted by Baseball Crank at 8:52 AM | Baseball 2005 | Comments (18) | TrackBack (2)

Excellent post. Regarding statistical evidence of the effect of steriods on Bonds' late-30s performance, take a look at this chart I put together for our blog on the Washington Nationals:

I agree that it we can't use stats to definitively prove that his use has affected his output, but, to me, this comes pretty close, and is much more relevant than Carroll's examples, for the reasons you state.

I have been surprised by many statheads response to the Bonds issue. For those who have rightfully championed the use of stats for objective analysis of the game, they seem to be denying what those stats tell them about this issue.

Posted by: DM at January 4, 2005 9:55 AM


Don't forget that pitchers take steroids too. Have to wonder when a journeyman reliever finds 5 to 7 mph on his fastball. Or how Clemens got so thick. His "growth" is almost as staggering as Bonds'.

Posted by: JT at January 4, 2005 10:34 AM

It's pretty obvious that Bonds has done either somthing Herculean or he had some help. The saddest part of all of this is he has given up many years of his life to get a record that will be forever tarnished.

My son is a college baseball player and MLB needs to send a strong message to him and his teamates about the cost of steroid use. Both the punishment and humiliation of being stripped of your records and the long term effects on your health.

I hope that the commissioner and the players union will make a stand but I think it will be up to President Bush to lead them in the right direction.

Posted by: roux at January 4, 2005 1:42 PM

Actually, force = mass x acceleration

momentum is m x v.

But the issue is the same, if you can get faster in the same space, you're accelerating more.

Posted by: Phil at January 4, 2005 1:56 PM

You haven't completely talked me out of the Aaron comparison yet. Or at least, I think it's more interesting than you're giving it credit for.

In 1999, the year he turned 35, Barry Bonds kicked up his home-run rate to one every 10.5 at bats, up from 14.9 the year before. The National League, meanwhile, kicked up its home run rate from one every 34.6 at bats, to one every 30.8. For the amount of at bats it took an average National Leaguer to hit 1 HR, Barry would hit 2.95, giving him a "Steroid Score" of 295.

In 1969, the year he turned 35, Hank Aaron kicked up his home-run rate to one ever 12.4 AB, up from one every 20.9 the year before. The National League, meanwhile, kicked up its home run rate from one every 61.6 AB, to one every 44.7. For the amount of at bats it took an average National Leaguer to hit 1 HR, Henry would hit 3.60, giving him a "Steroid Score" of 360.

To sum up: The seasons they turned 35, Aaron's contextual slugging was more impressive. Steroid Scores:
Bonds: 295
Aaron: 360

The Steroid Scores for when they turned 36:
'00 Bonds: 301
'70 Aaron: 291

'01 Bonds: 458
'71 Aaron: 454

'02 Bonds: 384
'72 Aaron: 352

'03 Bonds: 376
'73 Aaron: 435

Yes, Aaron was helped by his home park (hitting 25 more at home for those 5 years, though 16 where in '71). Yes, he was missed 30 and 40 games those last two years. But his spike in power -- as measured by how much more frequently he hit home runs than his peers -- is weirdly similar to Bonds'. There are other measurements, of course.

Posted by: Matt Welch at January 4, 2005 4:23 PM

Matt - Interesting take on the numbers; I'll take a closer look at them when I have the time. A few thoughts:

1. I still think Bonds' slugging percentage, not his home run rate, is the relevant variable.

2. Aaron's 1968 season is a bad reference point, given the offensive conditions in 1968.

3. Aaron had been a much bigger home run threat in his twenties than Bonds, esp. relative to the league.

4. I don't think you can write off Aaron's home field advantage and declining playing time that easily. Aaron's home slugging percentage compared to the road was +26 points in 1969, +19 in 1970, +240 in 1971, +3 in 1972, +75 in 1973, and +93 in 1974. That's huge.

Posted by: The Crank at January 4, 2005 5:10 PM

Seems to me Matt forgot to point out one very large factor - they lowered the pitcching mound height from 15 inches to 10 after the 1968 season.

Now on the other hand, you might have to also figure out how to factor in the TRUE size of the strike zone, which nowadays sems to be half of what it was in Aaron's day.

Posted by: Dave at January 5, 2005 12:51 AM

Just look at Bond's in the early 90's and now? I think Bond's Homeruns during the last few years were also affected by the Walks he gets. His homeruns could be even more if they would pitch to him. Bond is also hitting in a park that is not that good of a hitting park. Look at what Helton does in SF. Not very good. But Bond's hit's them out all the time.

Posted by: Ernie at January 5, 2005 12:16 PM

Thanks for taking the time with this issue. I'll agree that the Times piece doesn't do a great job of making the case. It was a limitation of space and editing. I'd much prefer that people discuss the BP piece that led to the Times asking me to do the op-ed. I feel it's more fleshed out.

That said, I offered Clay's examples as just the slightest evidence (as Matt's excellent stats also show) that there are reasonable explanations aside from steroids. I'm no statistician, but think the examples are valid. As players age, they tend to slow down. A guy like Stargell, never a speed demon, could still trot well, but maybe not leg out a double in the gap. Realizing this, maybe he swung for the fences a bit more. Wish I could ask him (and yes, trying to get hold of Aaron.)

Yes, I'm agnostic on steroids to the point that I'd like to see someone have a positive test under a valid testing system before we hold them up as criminal. I think steroids *are* a problem, *do* have an impact on the game, and *have* to be addressed by MLB/MLBPA. If I didn't, I wouldn't have written two pieces and sure wouldn't be doing a whole book on it (due out this spring.)

Posted by: Will at January 5, 2005 2:12 PM

DM is right. I find it ironic that the "stathead" community, the self-proclaimed seekers of "objective" truth in baseball, have suddenly lost their interest in finding out the truth when it comes to steroids.

Posted by: tc at January 6, 2005 9:10 AM

Reaction time has been neglected here.

There are plenty of guys who can bench press a mobile home who can't hit a ball.

Steroids improve one's reaction time. That is a far greater advantage than the obvious strength gains.

Note how quickly Bonds gets around on the inside fastball now. Formerly a weakness, he now crushes that pitch regularly. That's all reaction time.

Posted by: Zufall at January 6, 2005 1:38 PM

"Steroids improve one's reaction time"
Really? Hadn't heard that before. Got links?

Posted by: huski at January 7, 2005 4:13 AM

On the issue of the amount of rest that Aaron, Stargell, and Bonds got late in their careers:
I'll be slightly facetious here, but...
It is clear that while Bonds played a high percentage of games the last few years, it is also clear that he has gotten plenty of rest - on the basepaths, in left field, and at the plate while drawing so many unhittable pitches.
Maybe the large amount of time that he spends at half speed allows him to really focus those 2 times per game that it matters.

Posted by: Jeff at January 7, 2005 11:14 AM

steroid use is all bad and people shouldnt use it and sutff like that you know. barry bond use to be a idol to me and now i look up to a sorry player who cant hit for crap later

Posted by: dantheman at February 4, 2005 3:38 PM

I recently purchased the "juiced" book by jose, and all i heard out of it was snitching, respect for the steriod buisness, a sympathy story for jose, and an outright ripoff to sell a low rate book. I've never seen such a bad case of jealousy towards Mcgwire. Conseco obviously hates Mcgwire for stealing his show, and taking all of consecos fame and minimizing it into a plie of $&!#. He wants to take Mcgwirw down after his own life of doping and injecting himself with steroids, just to get 38 homeruns away from the 500 club.

Posted by: burge104 at April 12, 2005 6:06 PM

For my quarter project for my statistics class I did a statistical study on the top 20 home run leaders on all time. My claim was that if a player has a high home run average after the age of 32 then before the age of 32 that I would claim that they used steroids. I did a 2 - Sample T-Test for each player using and alpha level of .05. My conclutions showed that the only players that has statistical evidence at the .05 level were Barry Bonds, Sammy Sosa, Rafeal Palmero and Mark McGwire.
My stats teacher was very surprized and I got an A on my paper. Does anyone support my claim or have any questions/ suggestions?

Posted by: James F at May 23, 2005 1:27 PM

dick rollins is bad like steroids i think steroids should not be allowed. i can not express what i want to for i do not take english class any more

Posted by: little nik at May 24, 2005 11:19 AM


Posted by: JAMIE at September 12, 2006 2:56 PM
Site Meter 250wde_2004WeblogAwards_BestSports.jpg