Thursday, December 26, 2013

Free Preview: "Sports Drugs and Baseball Stats"

This is a preview of the Joe Sheehan Baseball Newsletter, an e-mail newsletter about all things baseball, featuring analysis and opinion about the game on and off the field from the perspective of the informed outsider. Joe Sheehan is a founding member of Baseball Prospectus, a contributor to Sports Illustrated and a featured guest on radio and television. He has been writing about baseball for nearly 20 years.

Your subscription gets you the newsletter and various related features three to five days a week, more than 200 mailings a year full of smart, fun baseball writing that you can't find in the mainstream. You also get access to the complete newsletter archive, more than 500 issues deep, of great baseball content.

You can subscribe to the newsletter for one year for $29.95, or for six months for $16.95. Use the PayPal button and the drop-down menus to pay using your PayPal account or major credit card.


The Joe Sheehan Newsletter
Vol. IV, No. 133
December 30, 2012

I have largely kept my discussion of the 2013 Hall of Fame ballot and the related issues to Twitter. This year's ballot, loaded with the villains of the era, has been coming for a while, and we're of course familiar with how sports-drug usage has affected the vote totals of Mark McGwire and Rafael Palmeiro in their time on the ballot. The number of problems with Hall of Fame voting in the early days of the 21st century start with the way the voters seem to be folding this particular brand of bad behavior into their evaluations, but they extend beyond that to the issues of narrative versus fact, and the way things like the 5% rule and the 15-year eligibility queer relative results, and the Hall's insistence on retaining the concept of the Veterans Committee decades past its useful life.

An exhaustive examination of many of these issues -- a couple are, admittedly, my own hobby horses -- would take a year, and I'm not signed on to write the update to The Politics of Glory, so I won't be tackling them all. Today, though, I want to focus on one particular point that undergirds the decision by many writers to withhold votes for some of the greatest players in baseball history. There's a belief among many writers and fans that the use of sports drugs with the intent of gaining strength queered the statistical output of baseball players during what is popularly known as "the Steroid Era". Specifically, it's a belief sports-drug usage caused a peak in overall home runs hit and a peak in the individual leaderboards. This is the distinction that separates strength-enhancing sports drugs from other forms of cheating, most specifically amphetamines.

This belief provides a veneer of statistical viability to a set of opinions that isn't backed by even 200-level analysis. It's an argument that requires you to ignore the majority of data points and nearly all of the factors that contributed to the spikes. What I find most interesting, though, is that it requires you to ignore the statistical evidence of amphetamine use that is exactly in parallel to what believers see as the statistical evidence of steroid use.

Labeling anything "the amphetamine era" is problematic, but we can say with confidence that amps were a big part of the game from the 1960s -- an era described by Jim Bouton in Ball Four -- through the 2000s, when testing for amps was put in place, with first counseling and then suspensions the punishment for testing positive. How deep the penetration of amphetamines, we'll probably never know, but we do know that some of the greatest players of the era have said that they did use, if perhaps only on occasion. (Remember that a single positive test for steroids has been enough to end Rafael Palmeiro's chance at the Hall of Fame.)

Historically, stolen bases were at their peak usage in the dead-ball era that ran through World War I. No season since 1919 has featured more than .9 stolen bases per game. Since 1920, however, there was one era in baseball history in which stolen bases were more prevalent than any other:

Peak SB/game, since 1920:

1987    0.85
1976    0.79
1983    0.79
1986    0.79
1988    0.79
1980    0.78
1990    0.78
1992    0.77
1982    0.75
1985    0.74

Were I to go past ten, you would find that there has been just one season since 1920 -- 1920 itself -- outside of the amphetamine era in which there were at least .7 SB/G. The league stolen-base totals in the "amp era" are as distorted as the league home-run totals were in the "steroid era". This extends to the individual leaderboards as well. There have been 23 seasons since 1901 in which one player has stolen at least 80 bases. Eighteen of those (78%) occurred from 1962 through 1988 -- the other five all occurred from 1910 through 1915.

Using the same "logic" that underpins the idea that steroids caused a bunch of home runs and caused a bunch of 60-homer seasons leads to the conclusion that amphetamines caused a bunch of stolen bases and a bunch of 80-steal seasons. Let me be clear: this is not my argument. It is, however, the counter to the wildly prevalent ideas that steroids caused home runs and were somehow different than amphetamines. You cannot look at the data and hold both of those ideas in your head.

Let's run at this from a slightly different direction. There have been 14 instances in MLB history of a player playing in at least 700 consecutive games. Six, including four of the top six, occurred during the amphetamine era. That's a small enough data set that it may not be meaningful, so let's look at something else. Baseball went to a 162-game schedule in two stages, the AL in 1961, the NL in 1962. From 1961 through 1972, there were a total of 20 instances in which an individual player played in at least 162 games -- a bit less than two a season. Here are the totals starting in 1973, through the end of the 26-team era:

1973: 6
1974: 5
1975: 3
1976: 3
1977: 4
1978: 5
1979: 10
1980: 9
1981: 0 (strike)
1982: 8
1983: 4
1984: 5
1985: 5
1986: 6
1987: 2
1988: 2
1989: 7
1990: 2
1991: 3
1992: 4

The widespread use of amphetamines certainly seemed to make it more likely that players would play in every game, with a clear peak in the 1979-82 period, broken up by the strike-shattered season of 1981. Once expansion pushed the industry to 28 and then 30 teams, the number of players playing in 162 games settled in at around five a year, with occasional spikes above that number through 2008. What is very interesting is that in the last four seasons -- since testing positive for amphetamines a single time became punishable by a 25-game suspension -- the average has dipped back to the 1960s figure of two a year -- just eight players have played in 162 games since 2009. In 2007 alone, seven players did.

In the early days of baseball, individual pitchers routinely through nearly all of their team's innings pitched. When "pitching" went from just that to more of a skill, individual innings pitched totals dropped, and they dropped further as pitchers threw harder spun the ball in an effort to miss the bats of hitters swinging as hard as they could. There have been 371 seasons of at least 300 innings pitched since 1901. They break down, by decade, like this:

1901-10: 142
1911-20: 89
1921-30: 34
1931-40: 19
1941-50: 13
1951-60: 8
1961-70: 29
1971-80: 37
1981- the end of time: 0

There were 40 300-inning seasons from 1931-1960. There were 37 from 1971-1980. There has not been one since then. The period from 1963 through 1987 also was the peak for reliever usage; 130 of the 166 seasons in MLB history in which a pitcher threw 120 games while making 90% of their appearances as a reliever occurred in this 25-year span. No pitcher has done this since 1991.

The amphetamine era featured just as many statistical anomalies as did the steroid era, but there was no connection between the two reported. No one cared. Why that is the case is a topic for a book, I'd imagine, but you cannot defend the idea that steroids alone fundamentally changed the game's statistics in a way that the previous generation's drug of choice didn't. The correct answer, or course, is to see the whole board and acknowledge that peak homer was the product of a dozen factors, with the number "73" an explicable statistical outlier in that context, just as peak steal (and "130") were the same, just as peak playing time was the same, just as peak innings pitched was the same.

We build assumptions into the language. Steroids are "performance enhancing" and amphetamines -- or cocaine -- are not, for no reason other than that's just the way the language developed, borrowed by and large from Olympic sports that have little to to with the practice of a major-league baseball season or an individual baseball game. This is important in the context of Hall of Fame voting because the Baseball Writers Association of America has already honored many men who used sports drugs. The only way to argue that they have not is to define amphetamines and steroids as dramatically different, and there is simply no rational case for doing so, not based on legality, not based on notions of cheating, and certainly not based on the populist approach to seeing drug use in statistical lines. If you see steroids here, then you should see amphetamines here.

The cadre of writers who see themselves as defending the Hall against cheaters and drug users who distorted the game is wrong. It's not a point of debate, it's not "reasonable minds can disagree", it's not a moral stance. These people are wrong. MLB turned a blind eye to drug use of all kinds for decades, as did the writers covering the game. Those same writers elected player after player after player from an era in which we know amphetamine use was popular without ever raising the issue of illegal drug use in the context anyone's candidacy. To draw the line at steroids as if they represent a bridge too far is ignorant of history, ignorant of statistics, and ignorant of baseball.