April 25, 2008
Lies, Damned Lies
The Age of Suspicion
Every time I do a chat aroud this time of year, when small sample sizes encourage people to leap to conclusions, I get a couple of questions like these (names have been changed to protect those who have lost their innocence):
Tony (Somerville): Ortiz looks MUCH smaller, though still just as fat, compared to 2007. His numbers are also down and at this point he's more likely to be voted Mr. Universe than get an extra-base hit. Why aren't there steroid rumors surrounding him? He is the perfect candidate for a roider, the power came out of nowhere, and after the Mitchell Report the power disappeared like cup cakes at a pot party.
I hate, hate, hate questions like these, and the mindset that underlies them. One of the beautiful things about baseball is its unpredictability. For as good as we might ever make PECOTA, for as much as we might think we have it all figured out, there are always going to be things we couldn't anticipate: a marginal pitching prospect becoming the best reliever in baseball, or a sure-fire, future Hall of Famer falling flat on his face in the middle of his career. Indeed, one of the reasons I created PECOTA was to account for this natural variability in performance.
Recently, there has become a tendency to ascribe any unattributed change in performance--favorable or unfavorable--to steroid use. Invariably, this is backed up with some sort of hearsay, or circumstantial evidence: "Ortiz looks much smaller this year" or "I hear so-and-so and Jose Canseco went to strip clubs together in Miami." Our brains are designed to look for patterns in static, and like it or not, there is a lot of static in baseball statistics. Steroids have become the cure-all for every bit of our cognitive dissonance.
So I was disappointed recently when I saw that Bill James, a man whose work I greatly admire, made exactly these sorts of arguments in a comment in the Bill James Gold Mine 2008 about the mid-'80s Minnesota Twins. I've sat on this subject for several weeks, as it set off a small conflagration in the blogosphere that quickly burned itself out, as I have some trepidation about launching friendly fire at someone whom I consider a colleague in spirit. But at the end of the day, I think those of us in the analytical community have a responsibility to let the facts speak for themselves. So, here goes.
The article that touched off the controversy was posted at the Bugs and Cranks blog, and included the following passage from James:
Two of the greatest home run under-producers of all time were teammates: Kirby Puckett and Gary Gaetti in 1984. Puckett hit no home runs (-16), Gaetti hit only 5 (-19). Suggesting the possibility that the Twins' two World Championships may have been aided by their team being among the first to discover…well, I'd better not go there. Nor will I point out that Gaetti was bald and had acne and Puckett died young.
It has been suggested that this comment was snark, but, if so, I guess I don't get the joke. I think the steroids controversy has been vastly overplayed by the media--ironically, perhaps, because of our collective reverence for baseball's history and statistics--and has more to do with baseball past than baseball present. I like to remind people that none of the players who received even a single vote for MVP, the Cy Young, or the Rookie of the Year Award last season appeared in the Mitchell Commission report, and none of them has otherwise been the subject of a credible claim about PED usage. Nevertheless, accusations of steroid use are taken awfully seriously by an awful lot of people--and players who are the subject of such accusations stand to lose millions of dollars worth of contract and endorsement money, not to mention a lifetime's worth of reputation.
Notwithstanding that we shouldn't feel the compulsion to murder by dissecting any time we see something out of the ordinary in baseball, one of the ironies is that both Gaetti's and Puckett's biographies seem to offer a multitude of alternate hypotheses. James's implication about Gaetti, particularly, seems like a stretch, as Bugs and Cranks does a pretty good job of pointing out:
Meanwhile, his jab at Gaetti is based on…baldness and acne? Sure, Gaetti's home-run total in 1984 (five) was 19 below his lifetime average. But 1984 wasn't his rookie year (like it was for Puckett). Gaetti came up in 1982 and hit 25 home runs that year, followed by 21 in 1983. After his one-year power shortage in '84, he hit 20, 34, 32, 28, 19, 16, 18, 12, 14, 12, 35, 23, 17, 19, 9 and then 0. So what happened? Did Gaetti decide to stop doing 'roids for the '84 season, and then, unhappy with the results, decide to resume use in '85?
Gaetti was an 11th overall pick in the draft who hit for power from the very moment he put on a professional uniform. He belted 14 home runs in just 66 games in his rookie season at Elizabethtown, and then 22 more at Wisconsin Rapids, and then 30 in Orlando; the next year, as a rookie with the Twins, he hit 25 in the major leagues. Gaetti's 1984 campaign, the "atypical season" that James refers to, was the only time in his career when he played a full season and hit fewer than a dozen home runs.
So what happened that year? According to contemporaneous accounts in the New York Times, Gaetti, to that point a lifetime .237 hitter in the major leagues, was trying to lower his strikeout rate and improve his batting average, which he succeeded in doing, cutting his strikeouts by 50 percent and hitting a temperate .262. But this came at the cost of a significant amount of power. You can almost see what happened by looking at his monthly splits: Gaetti hit .310 in April, and was no doubt convinced that he was doing the right thing--but this is a story about how a player who gets the wrong message about which statistics are important can wind up doing tangible damage to his club, and not a story about steroid use.
Make no mistake, players today are getting smarter about this stuff. Mark Reynolds, a player whose numbers bear a strong resemblance to Gaetti's, went into spring training this year with a plan of cutting down his swing and reducing his strikeout total. He abandoned that plan in mid-spring--and a good thing, too, as he has been rewarded with six home runs and a 912 OPS, rather than suffering through a Gaetti '84 type of year.
For Puckett, the power was a little slower to come. Unlike Gaetti, he hadn't hit for much power in the minor leagues, and then he hit just four home runs over his first two seasons in the majors, a period that comprised more than 1200 at-bats. He also hit for (empty) .296 and .288 averages in his first two seasons in the majors, and reached double digits in stolen bases, and banged out a lot of triples. It is plausible that, if Billy Gardner's Twins were giving a born power hitter like Gaetti the wrong message about just which of his skills were important, they were certainly doing the same with a player like Puckett, who might have looked to them like their own version of Tony Gwynn.
However, Puckett was always someone who was regarded as having a lot of potential, even if his skills were a little unrefined. Like Gaetti, Puckett had also been a high draft pick; he was taken third overall in the January phrase of the 1982 amateur draft. While the players selected in the (long-since defunct) January draft weren't usually as talented as their June counterparts, the fact is that there were 338 players selected in that draft, and Puckett was selected third from that group. Obviously, somebody saw something in him, and it wasn't because of some impeccable baseball pedigree; on the contrary, Puckett was unnoticed out of high school and took a job at an auto plant before eventually attending Bradley University and letting the world discover his gift for baseball.
Everything clicked for Puckett in 1986 when he worked with Tony Olivo in spring training on driving the ball, something which came naturally enough for him. PECOTA has found that a player's weight, rather than his height, bears a far stronger predictive relationship with his power development, and a player who tips the scales at Puckett's 210 pounds will have no lack of momentum on his swing. So, Puckett was a little bit of a late bloomer. As a baseball player (not as a human being) he's a little bit reminiscent of Curtis Granderson, another Chicago native who took a bit of a relatively obscure path to the major leagues, and who remains the only major leaguer in history to have played his college ball at the University of Illinois at Chicago.
Was Puckett's path to becoming a power hitter unusual? Yes. But it also wasn't completely out of context, and it isn't without explanation, any more than it would be any time that a player parlays a combination of prodigious natural talent, hard work, and great instruction and becomes something more than he used to be. Definitionally, the career paths of remarkable players are unique and special, and if we make the mistake of trying to frame their accomplishments in terms of what ordinary players might do, we may find that we have been misled. There are 750 human beings in the major leagues at any given time, and thousands of human beings in professional baseball. Some of them are going to behave like human beings, and begin to do things that they hadn't done before. If we are so cynical that we must ascribe any such performances to steroid use or some other sort of ulterior motive, we might as well have PECOTA simulate everything and call it a season.