Archive for Steroids
Steroids and Expansion
In a recent post, I briefly discussed the possibility of expansion being a contributor to the rise in home runs. It’s based on the work of Stephen Jay Gould, who posited that as talent became less (more) disperse, excellent achievements became less (more) likely to occur. Since home run hitters tend to be baseball’s best hitters, there improvement should be expected as pitching talent becomes more disperse. Expansion is one cause of talent dispersion.
Reader Shek sent me this excellent graph (actually, he sent me his beautifully crisp Stata code to recreate it) of home run rates and the number of teams per season.
The rise in home runs does seem to move with expansion in the 1990s and possibly the late-1960s, but the relationship is hardly airtight, nor necessarily causal. And even in the 1990s, it’s difficult to know if the corresponding spikes are noise or real effects.
But, expansion isn’t the only cause of dispersion. In Chapter 8 of The Baseball Economist, I discuss how league dispersion has changed over time, by measuring the variation in performance across players. With and without expansion, the difference between the best and worst for pitchers and hitters has fluctuated quite a bit with time, as the graph below reports.
The 1990s and 2000s are two of the most disperse decades for pitchers in the history of baseball, while there were many good pitchers who excelled, there were also many bad pitchers for batters to feast on. However, during other past expansions, pitching quality was relatively compact. So, it should be no surprise that recent expansions felt more of an effect than past expansions. But still, this is not proof, just evidence that fits with a theory that is very difficult to test.
Thanks to Shek for passing the graph along.
What Caused the “Steroid” Era?
With Mark McGwire’s recent admission that he used steroids throughout the 1990s, there has been a resurgence of chatter regarding how to view the performances of the past 15–20 years. But, I think it’s wrong to attribute most of the high home run rates of this time-frame to steroid use. The connection that many people want to make is understandable. We have evidence that many players were using steroids during this era, and steroids have been shown to increase strength and power; therefore, it should follow that home run rates should rise accordingly. (And let me make this clear before I move to the next step: I believe wholeheartedly that steroids are effective ergogenic aids to baseball players—it’s growth hormone that does nothing.) However, I do not think steroids are the main cause of the dramatic rise in offense (particularly the home run) in the 1990s–2000s.
Why do I think this? Take a look at the following graphs. The first maps home runs per game from 1921 — 2009.
It’s clear that the present era is different from the past, but I also think it’s interesting how quickly it changed and did not change. In 1993, home runs per game jumped by 23%, and in 1994 they jumped 16%. Since that time, the average absolute change in home run rates has averaged about 5% and maxed out at 9%. Home run rates rose, then plateaued. The graph below zeros in on home runs per game since the 1990s.
If steroids were the cause of the steroid era, then we should have gradually seen them enter the game. A few players use and then others slowly adopt their technique. But that’s not what we observe. Almost overnight, home runs jumped. If you want to believe home runs are largely responsible for the change then you have to believe that players all got together in 1992 and 1993 and said, “hey, it’s juice time.”
But even more convincing in my mind is the fact that the home runs haven’t gone away with steroid testing. In 2005 (marked by a vertical line on each graph), Major League Baseball began drug testing with suspensions, and the home runs didn’t go away. You might be able to look at the graphs and identify a slight decline—it’s difficult to separate the noise from real changes with so few observations—but it’s clear that home runs aren’t close to their pre-1993 level. Yes, tests are imperfect and can be beaten, but the dramatic change in enforcement—from no monitoring or punishment to strict monitoring with punishment—ought to yield more substantial declines. Do I think steroid testing has taken away some power from its users? Absolutely, and it barely shows up in the aggregate data. In my mind, the rise must be attributable to something else. And if you think Ken Griffey had two seasons of 56 home runs without the help of steroids, then it would be useful to have an alternate theory.
So, if steroids aren’t the cause, then what is? I have a few theories.
1. MLB changed the ball. I have little doubt that the league allowed (or introduced) a lively ball into the league. This would have the effect of boosting home runs for all players. Certainly, I can’t prove this, but there have been whispers about it for years. This theory explains the dramatic rise in home runs and the lack of decline after testing.
2. Expansion. Baseball expanded by two teams in 1993 and two more in 1998. As I have detailed before (see here and here), as talent became diluted, excellent performances began to happen as the very best (hitters and pitchers) were able to take advantage of the very worst. The abrupt change in home runs fits exactly with the timing of expansion. Joe Posnanski also discussed the the possible influence of expansion (as well as a few other explanations) in a recent column.
3. New stadiums. Smaller parks and home-run alleys may have caused more home runs to fly out of the ball park. While, on balance several new stadiums are home-run-friendly (Colorado) others have dampened home runs (San Diego). Given the dramatic spike in home runs, I don’t believe this could have had more than a minor effect on home runs, and the effect would occur gradually.
4. Bats. Bat technology has certainly changed since I was a kid, when nearly every player used an ash bat made by Louisville Slugger. Maple bats, hardened with shellac, with tiny handles are the new weapon of choice for batters. Again, I think this fails to explain the home run spike in the early-1990s, because they trickled into the game. It could be a contributor, but it’s not sufficient on its own.
5. Strike zone. It’s possible that the league redefined the strike zone into a more-compact area to increase offense. However, recent attempts to expand the zone (with the help of computer monitoring) haven’t dampened offense.
There may be a few explanations that I have missed, and I’m open to others. But if you have another hypothesis, it must be able to explain the near-instant rise and subsequent plateauing of home run rates. I believe steroids fail this test miserably. I’m fine with the “home run era,” but in my mind steroids can only explain a very small part of why home runs increased during this time period.
Thoughts on Mark McGwire
Unfortunately, I’ve had a glut of things pass across my desk so I haven’t had the chance to read much about Mark McGwire’s admission that he used steroids. How do I feel about this? I’ll recycle an old post.
This will be short and simple.
For the third year in a row, Mark McGwire did not receive sufficient votes from the Baseball Writers Association of America (BBWAA) to be elected to the Hall of Fame. The explanation is simple: many writers feel that his performance was aided by performance-enhancing drugs. There are certainly several sources for accusations, but they have some credibility problems. Others point to his continued excellent performance into his thirties and his bulging biceps. McGwire fits the profile of a steroid user, and I wouldn’t be surprised if he admitted to using them. McGwire hasn’t helped his case by refusing to testify under oath before Congress; however, that is the advice that every lawyer would give to his client in these circumstances.
I don’t want to pick a fight regarding whether or not he used steroids. I don’t really care. My argument is simple. Let’s assume McGwire used hard-core anabolic steroids every day of his baseball career. He didn’t violate a single baseball rule. Mark McGwire played his last baseball game in 2001. It wasn’t until 2004 when anabolic steroids became a punishable offense despite the fact that serious doping regulations had been instituted in nearly all other sports. McGwire shouldn’t be excluded any more than any other player who drank amphetamine-laced coffee prior to its ban. He’s being barred from the Hall of Fame for doing something that people wish was against the rules but wasn’t.
File Under True, But Misleading
Rob Neyer points to an AP story about the “rise” in stimulant exemptions for ADHD among MLB players.
Stimulant exemptions in MLB slightly rise again
NEW YORK (AP) -The number of baseball players authorized to use otherwise banned stimulants for ADHD rose for the second straight year.
Baseball granted 108 therapeutic use exemptions for attention deficit hyperactivity disorder during the year ending with this World Series, according to a report released Tuesday by MLB’s independent drug-testing administrator. That was up from 106 a year earlier and 103 in 2007.
So, in three years total exemptions have risen by a grand total of five…and an increase in exemptions is the story? I’m more shocked that players aren’t flocking to known effective performance-enhancing drugs (unlike growth hormone) through a legal exemption. These are the same drugs that for years players openly mixed into clubhouse coffee pots.
Where Is the Caffeine Outrage?
Int J Sport Nutr Exerc Metab. 2008 Aug;18(4):412-29.
The effect of caffeine as an ergogenic aid in anaerobic exercise.
Woolf K, Bidwell WK, Carlson AG.
Department of Nutrition, Arizona State UniversityThe study examined caffeine (5 mg/kg body weight) vs. placebo during anaerobic exercise. Eighteen male athletes (24.1+/-5.8 yr; BMI 26.4+/-2.2 kg/m2) completed a leg press, chest press, and Wingate test. During the caffeine trial, more total weight was lifted with the chest press, and a greater peak power was obtained during the Wingate test. No differences were observed between treatments for the leg press and average power, minimum power, and power drop (Wingate test). There was a significant treatment main effect found for postexercise glucose and insulin concentrations; higher concentrations were found in the caffeine trial. A significant interaction effect (treatment and time) was found for cortisol and glucose concentrations; both increased with caffeine and decreased with placebo. Postexercise systolic blood pressure was significantly higher during the caffeine trial. No differences were found between treatments for serum free-fatty-acid concentrations, plasma lactate concentrations, serum cortisol concentrations, heart rate, and rating of perceived exertion. Thus, a moderate dose of caffeine resulted in more total weight lifted for the chest press and a greater peak power attained during the Wingate test in competitive athletes.
Nice Work, Federal Government!
A federal judge has barred prosecutors from using three positive steroid tests and other key evidence in Barry Bonds’ trial next month because of his personal trainer’s refusal to testify.
Federal agents seized the samples when they raided BALCO in 2003.
All that work for nothing. And when A-Rod starts suing to find the leakers, it’s going to get even more expensive. What an absolute embarrassment.
More on A-Rod’s “Power Surge” in Texas
Yesterday, I was watching ESPN in the wake of Alex Rodriguez’s press conference when some numbers flashed on the screen. It listed at Alex Rodriguez’s isolated power (SLG-AVG) in Texas and New York, showing a much larger number with the former, during the time that he admitted to using steroids. Though it was not explicitly stated, I guess viewers were supposed to infer that this was evidence of A-Rod’s steroid use; or, at least, I can see how someone might interpret these numbers as such.
I previously addressed A-Rod’s power surge in Texas using his home-run rate, and I found the gain was about one home run per season, which is too small to indicate a meaningful change. As I stated in my previous analysis, two problems with using Rodriguez’s raw numbers are that they don’t account for park effects and aging. But, homers are not the only measure of power.
When we go beyond home runs, the analysis changes slightly. The nice thing about aging and home-run rates is that they peak at around age 30, while doubles and triples peak at around age 28. This means that when we compare A-Rod’s extra-base performance in New York to Texas, we are going to expect a decline in doubles and triples from natural aging as well as from moving to a tougher park. In terms of home runs, A-Rod was still on the upswing when he left Texas.
The table below lists A-Rod’s career performance neutralized for park and era effects—these are not his actual numbers, which are polluted by ballpark and era effects—as well as a correction for natural aging. The “Aging (v. Peak)” column reports the percent difference from his projected peak doubles-plus-triples rate (DPT/AB). The aging estimates come from my forthcoming paper on aging in baseball. I base the aging progression towards his peak using the mean of his age 23 and 24 performances to project his peak DPT/AB performance of 5.94% at age 28. This baseline appears to overestimate his future DPT/AB a bit, but all I can do is make an estimate. The DPT/AB projection has the greatest variance fluctuations among the metrics I used to estimate the aging functions of hitters; thus, it’s not surprising to see them fluctuate as they do. However, what the numbers indicate is that his non-homer extra-base hits don’t appear to be vastly different that we would expect given the changes in park and aging, because even if I lowered his expected DPT/AB it would not show a spike during the Texas years.
Year Age Neutral Aging Pred. Neutral Pred. Neutral - DPT/AB (v.Peak)DPT/AB DPT DPT Pred, 1994 18 0.00% -34.15% 3.91% 1995 19 5.10% -27.81% 4.29% 1996 20 8.70% -22.12% 4.63% 1997 21 7.19% -17.08% 4.93% 1998 22 5.83% -12.69% 5.19% 1999 23 4.67% -8.95% 5.41% 2000 24 6.51% -5.86% 5.59% 2001 25 5.41% -3.42% 5.74% 34 36.03 -2.03 2002 26 4.54% -1.63% 5.84% 28 36.05 -8.05 2003 27 5.85% -0.49% 5.91% 35 35.35 -0.35 2004 28 4.32% 0.00% 5.94% Gain 01-03 -10.43 2005 29 5.19% -0.16% 5.93% Gain 01-02 -10.08 2006 30 4.59% -0.97% 5.88% 2007 31 5.45% -2.43% 5.80% 2008 32 6.45% -4.53% 5.67%
One thing to note is that a decrease in doubles and triples could mean an increase in power as a result of those extra-base hits becoming home runs. However, there doesn’t appear to be much of a spike in home runs.
The usual caveats apply to this kind of analysis. What if he’s lying? Well, that would change things, but all I can do is test out his story. What about his admitted use of Ripped Fuel (contained the stimulant ephedra, and was legal) before going to Texas? Again, there is not much I can do here. My aging estimates are based on the past when amphetamine use was widespread, so the aging function may not be steep enough to capture his expected decline, and we can just acknowledge that. What if I used a different time to estimate projected peak age? Yes, that could have an effect, it could go down or up.
The point here is that there is not much to see in these numbers. The so-called “power surge” that we saw in Texas wasn’t all that extraordinary given the park change and aging. And before anyone freaks out, please know that I believe anabolic steroids improve athletic performance, and they ought to help baseball players hit, pitch, run, etc. That there doesn’t appear to be an obvious boost in one player’s numbers doesn’t mean all that much in the grand scheme of things. But, I think it would be a worse sin to look at A-Rod’s numbers and suggest there was a significant power boost from steroids.
Two Topics I Like to Write About
Cal Warlick continues his excellent cartoon work in the AJC.
Wow, the Gwinnett stadium and steroids. How could I not link to this? 
Random Thoughts on Steroids
— Roy Oswalt understands the the importance of relative competition for measuring the productivity of baseball players.
“The ones that have come out and admitted it, and are proven guilty, [their numbers] should not count. I’ve been cheated out of the game,” Oswalt continued.
As a Ranger, Rodriguez was 3-for-5 vs. Oswalt with two doubles, one home run, three RBIs and two walks. Last year, as a Yankee, Rodriguez was hitless in two at-bats against Oswalt.
“The few times we played them, when he got hits, it could have cost me a game,” Oswalt said. “It could have cost me money in my contract. He cheated me out of the game and I take it personally, because I’ve never done [PEDs], haven’t done it, and they’re cheating me out of the game.”
— Former MLBPA head Marvin Miller notes that ignorant coverage of performance-enhancing drugs’ benefits may actually encourage use.
“A kid who would love to be a professional athlete reads the sports pages or watches ESPN and is told over and over again, ‘These are performance-enhancing drugs. They will make you a Barry Bonds or an A-Rod or a Roger Clemens.’ The media, without evidence, keep telling young people all over the country, ‘All you have to do to be a famous athlete with lots of money is take steroids.’ The media are the greatest merchants of encouraging this that I’ve ever seen.”
Miller is a bit over the top here, because I think that anabolic steroids likely do enhance performance and players should want some testing system to prevent their use; however, when it comes to growth hormone, he’s dead on. The media has completely botched the coverage of this issue, which is one of the reasons why I think that growth hormone should be legalized.
The point is that players have a strong incentive to gain an edge on each other. This road will inevitably lead many of them to seek out illicit solutions in an area where the experts are the guys who sell the stuff. And when they investigate further, they find prominent sports reporters declaring that HGH is just as effective as steroids. Do you think players are going to search through the scientific literature on PubMed? Heck, if I didn’t share an office suite with exercise physiologists, I probably wouldn’t know any better.
At the end of the day, players are going to take a long hard look at the list of prohibited substances. The fact that these drugs are banned will be sufficient to convince most players that the performance-enhancing benefits are real.
— Here is my proposed performance-enhancing drug policy for baseball in the NY Times, with further explanation here.
First, I suggest a system of fines and bonus. This is a Pigouvian tax and subsidy system that taxes players in accordance with the external costs that users impose on non-users—users may feel the personal benefits of a higher salary outweigh the health risks—and then transfers the financial gains to non-users who earn relatively less due to the fact that they chose to remain clean. This has the deterrence effect similar to suspensions; however, the substantial fine revenue gives players who feel they are in a use-or-lose situation an incentive not to use and to identify new cheating methods.
Second, I propose handing over all monitoring and testing to the players. It is the players who suffer the most from steroids. They are in an arms races where steroids make no individual relatively better than any other player—hence, there is no financial gain—yet, users end up suffering health consequences. This resembles a prisoner’s dilemma game.
What Impact Did Steroids Have on Alex Rodriguez’s Home Run Performance?
Following yesterday’s post, I received an e-mail from an interested reporter asking me if it was possible to compare Alex Rodriguez’s performance versus expected performance during his admitted steroid seasons.
Something about the wording of his e-mail stirred me to think about a better way to look at this than what I had previously done. The procedure is simple, I use Baseball-Reference’s “Neutralize Stats” tool to convert A-Rod’s home-run rate (HR/AB) performance to a consistent baseline. Then I use my estimates of aging from a study to be published in Journal of Sports Sciences (working paper version) to examine Rodriguez’s neutral career aging trajectory.
The table below lists A-Rod’s career performance neutralized for park and era effects—these are not his actual numbers, which are polluted by ballpark and era effects—as well as a correction for natural aging. The “Aging (v. Peak)” column reports the percent difference from his projected peak HR/AB. I base the aging progression towards his peak using the mean of his age 23 and 24 performances to project his peak HR/AB performance of 9.18% at age 30. It’s interesting to note that A-Rod’s peak HR/AB performance occurs at age 31 at 9.37%.
Year Age Neutral Aging Pred. Neutral Pred. Neutral HR HR/AB (v.Peak)HR/AB HR HR - Pred. HR 1994 18 0.00% -67.63% 2.97% 1995 19 3.18% -56.73% 3.97% 1996 20 5.63% -46.79% 4.89% 1997 21 3.94% -37.80% 5.71% 1998 22 5.98% -29.77% 6.45% 1999 23 7.91% -22.70% 7.10% 2000 24 7.41% -16.58% 7.66% 2001 25 8.12% -11.43% 8.13% 51 51.07 -0.07 2002 26 8.91% -7.23% 8.52% 55 52.56 2.44 2003 27 7.53% -3.98% 8.82% 45 52.72 -7.72 2004 28 5.98% -1.70% 9.03% Gain 01-03 -5.35 2005 29 8.28% -0.37% 9.15% Gain 01-02 2.37 2006 30 6.00% 0.00% 9.18% 2007 31 9.37% -0.59% 9.13% 2008 32 6.84% -2.13% 8.99%
The reporter also noted that A-Rod claims to have quit taking steroids after a 2003 spring training injury; therefore, we might not want to include 2003 as a steroid season. Thus, I include his 2001–2003 and 2001–2002 total gains in home runs. As the table indicates, 2003—the season in which we know he tested positive—he hit nearly 8 fewer home runs than expected. He hit almost exactly as many home runs as expected in 2001 and 2.44 more than expected in 2002.
So, what were A-Rod’s steroids worth? 2.37 home runs over two seasons, or a little over one home run a season. At least, that is the estimate based on the method I laid out above; however, it’s probably best to say that there was no observed effect. It is possible that the steroids did give Rodriguez a boost, and this may have helped him through an injury or some other factor that my estimate does not account for. It’s also likely that he hit more home runs than expected through random chance. Given the general swings in the play of the game, it is very difficult to separate true performance changes from random swings in performance. The deviation here isn’t large enough to say much.
The important finding is that the statistical record doesn’t reveal an obvious spike in home-run performance by Alex Rodriguez during the time when he admits to using performance-enhancing drugs.
Addendum: The reporter mentioned above is Carl Bialik, who now has a post up on the subject at The Numbers Guy.
Further Addendum: For those who have asked why posted this brief analysis, see Thomas Boswell’s latest.
Rodriguez may have taken performance-enhancing drugs for only three years — never before, never after.
For one thing, his statistics, as we’ll show, indicate that he may be coming clean. He averaged 33 percent more homers in his dirty Texas years — from 2001 to 2003 — than in the other 10 full seasons of his career. That’s a huge leap, similar to the numbers that first incriminated Barry Bonds in many baseball minds….
In his three years in Texas, from 2001 to 2003, he averaged 52 homers vs. 39.2 everywhere else. The jump was even bigger when compared to his previous five superstar years in Seattle, when he averaged 36.8 homers.
After hitting 42, 42 and 41 homers in his last three years in Seattle, he hit 52, then 57 in his first two years in Texas. Granted, the Ballpark is a launching pad. Rodriguez slugged .666 there in three years vs. .576 on the road. But that’s still worth only a few extra homers a year, not 12 or 15.
