Melky Cabrera’s Positive Test, Not Performance, Proves PED Use

Melky Cabrera was one of baseball's most interesting feel good stories, that is, until Wednesday, when the San Francisco Giants' left fielder tested positive for increased levels of testosterone. Now, the once emerging cult hero has become a public enemy.

Not surprisingly, the response to Cabrera's positive drug test has been very emotional. Some have suggested that his statistics should be labeled with an asterisk, while others have demanded that he be disqualified from the batting race. Not content to only punish Cabrera, at least one pundit has suggested that the Giants should be forced to forfeit games in the standings. Although the outcry is understandable, most of the reaction has bordered on irrational. However, the most disturbing argument doesn't deal with an appropriate punishment. Since the revelation of Cabrera's positive test, some have suggested that the outfielder's improved performance was a smoking gun, which is particularly dangerous logic because it can be used to cast suspicion even in the absence of evidence.

Before the abrupt end to his season, Cabrera had boosted his OPS by 12% (OPS of .906; OPS+ of 157) compared to last year. Should that increase have set off alarm bells? There are 12,000 qualified seasons in fangraphs.com’s data base, but for the purpose of comparison, the first one of each player's career can be eliminated (because there is no prior point for comparison). That whittles the population down to 9,464, and among that sample, just under 14% of the seasons involved a greater improvement than the one Cabrera experienced in 2012. Although impressive, that doesn’t exactly qualify as abnormal.

Top-20 OPS Rate Improvements in Qualified Seasons, Since 1901

Note: Non-qualified seasons omitted for comparison purposes (i.e., all comps made to the most recent qualified season, even if a non-qualified season occurred in between).
Source: fangraphs.com

At this point, someone is probably saying, "What about 2011? I’ll bet Cabrera was using testosterone that year too". Presumably, if Cabrera was using PEDs last year, he would have been caught, but maybe his dumb luck just happened to run out this season? If so, that also calls his 2011 season into question. That year, Cabrera boosted his OPS by 21% to .809 (OPS+ increase from 83 to 121), which ranks within the top 4% of all performance spikes in qualified seasons since 1901. This same relative percentage holds true even when only comparing players in their fifth qualified season. Surely, such a significant increase now warrants a raised eyebrow?

Top-20 OPS Rate Improvements in Player’s Fifth Qualified Season, Since 1901

Note: Non-qualified seasons omitted for comparison purposes (i.e., all comps made to the most recent qualified season, even if a non-qualified season occurred in between).
Source: fangraphs.com

Although Cabrera’s younger age (just entering his prime in 2011) and depressed 2010 season (his 2011 OPS was only 7.5% higher than his 2009 rate with the Yankees) are mitigating factors, it’s not unfair to say Cabrera’s 2011 campaign with the Royals featured a historic unadjusted OPS increase. However, that still doesn’t mean his performance increase resulted from his use of testosterone. If one insists on making that connection, then what about the 418 other players who had OPS spikes as high or better than Cabrera’s? Did the Philadelphia Athletic’s Joe Dugan use PEDs to increase his OPS by 63% in 1920? What about Lou Gehrig, whose third qualified season featured an OPS spike of 28%? And, for Yankee fans quick to jump to conclusions, what do they think about Robinson Cano’s 22% OPS spike in 2009, which also happened to be his fifth qualified season?

Thanks to baseball’s improved drug testing program, we know that Melky Cabrera used performance enhance drugs. However, that doesn’t mean we know the extent to which his performance was enhanced. Although Cabrera’s offensive rise fits nicely with the narrative of PEDs being baseball’s equivalent of a magic potion, there are too many counter examples to contradict and not enough scientific data to support the inferred correlation.