Robinson Cano, Carlos Gomez, and the Illusion of Certainty

Words matter. In the context of the law, they can make all the difference.

They can also matter quite a bit in the context of PEDs and baseball.

Recently, Robinson Cano tested positive not for a steroid but rather for furosemide, a masking agent. This is how it was covered in the media.

ESPN:

A source familiar with the case told ESPN’s T.J. Quinn that Cano tested positive before the season and appealed. During the appeal, MLB apparently was able to determine his intent, which resulted in Cano dropping his appeal, the source said.

TJ Quinn himself:

Robinson Cano suspended for taking furosemide, a diuretic commonly used to mask PED use. It’s the kind of drug a player is likely to say he took by accident and didn’t help his performance. Eager to hear his explanation, because he has access to certified, clean supplements.

USA Today:

The IPA needed proof that Cano was using the drug as a masking agent.

The investigation revealed that Cano had clear intent to mask another illegal drug.

Cano was charged with a positive test, no different than if he were taking anabolic steroids.

That’s a loaded word, “proof.”

The trouble is that different kinds of proof exist. First, there’s the burden of proof. In other words, whose responsibility is it to prove their case? In civil law, it’s the plaintiff who has the burden of proving its case, and the defense has the burden of proving defenses. But even that is a bit misleading; the defense doesn’t have to prove anything. If the plaintiff doesn’t prove every legally required part of its case, the defense wins even if the defense provides no evidence at all.

Then there is the question of how much evidence a party need to prove its case. You’ve heard how much is needed in criminal cases if you’ve ever seen an episode of Law and Order: “Beyond a reasonable doubt.” That means the prosecutor must produce enough evidence that the finder of fact, be that a judge or jury, decides that the defendant’s guilt is the only reasonable conclusion. Note that the defendant’s guilt need not be the only possible conclusion. The typical academic formulation of this is that the defendant has a 95% chance of being guilty — which means, in slightly different phrasing, that if the defendant were tried 100 times, they’d be found guilty in 95 of them. This disparity between absolute proof and legal proof is why things like Alford and no contest pleas exist. So already the water is murkier. There’s more, though.

Civil trials, where only money is at stake, have much lower standards of proof — either “clear and convincing evidence” or “preponderance of the evidence.” The preponderance standard is simply that you must show it to be “more likely than not” — in other words, 50.1% likely. Clear and convincing is somewhere in between the preponderance and reasonable doubt-standards; for our purposes, you can think of it as around 67-75% likely.

There are a couple of standards of proof in the Major League Baseball Joint Drug Prevention and Treatment Program, which is incorporated into the CBA, that concern us. First, this is the relevant part, on page 21, about diuretics:

The presence of a Diuretic or Masking Agent in a Player’s urine specimen shall be treated as a positive test result if the IPA determines that the Player intended to avoid detection of his use of another Prohibited Substance.

But notice there’s no evidentiary standard listed. However, this is what the Agreement says about burdens of proof:

[T]he Commissioner’s Office shall have the burden of establishing that a Player’s test result was “positive” (as that term is defined therein), and that the test result was obtained pursuant to a test authorized under the Program and was conducted in accordance with the Collection Procedures and Testing Protocols of the Program and the protocols of the Montreal Laboratory (herein collectively “the Collection Procedures”). The Commissioner’s Office is not required to otherwise establish intent, fault, negligence or knowing use of a Prohibited Substance on the Player’s part.

And what about defenses?

So this is pretty much like a civil suit. The Commissioner has to prove the positive test by a preponderance, and a player may establish affirmative defenses (like a lack of fault or negligence) by a preponderance or mitigation (no significant fault or negligence) by clear and convincing evidence.

So why do we care? In this case, because MLB didn’t prove that Cano used PEDs beyond a reasonable doubt. They didn’t have to. All they had to prove was that, in 100 trials on the issue, they’d win 51 of them. If Cano needed clear and convincing evidence to prove mitigation, however, he’d have to prove that he had a 67% chance of winning — in other words, that he would win 67 out of 100 trials.

There are two consequences of this. The first is that Robinson Cano was not, contrary to media reports, proved to be a user of PEDs with 100% certainty.  Nor is it correct to state that MLB proved Cano used furosemide to mask PED use. Instead, what MLB proved was that it’s more likely than not that Cano used furosemide to mask PED use, which means something very different. And second, you can now see why Cano, faced with a lengthy absence from his injury, would elect to drop his appeal and accept the suspension. The numbers, after all, made his chances of winning remote, because he had to prove his case to a higher standard than MLB did.

But while we’re on the subject of PEDs, let’s look at another recent opinion. In the wake of the PED suspensions of Cano and Welington Castillo, Rays outfielder (and current disabled-list resident) Carlos Gomez had some thoughts.

“It’s not random. It’s not random. I can put my hand on fire, it’s not random,” Gomez told Yahoo Sports MLB Podcast host Jeff Passan. “They pick guys. I think it’s something the way you play, the way you act … I’m the oldest guy on the team. I get here earlier than everybody. Why? Because I have to work harder to maintain my body to support the rest of the season. When I do that and they come to you and have a drug test every time, you get furious. You get mad. One month into the season I got like seven drug tests. Something like that. Between five or seven. That’s not right. We have a guy on the team who for sure hasn’t had one drug test.”

Gomez told Yahoo sports that he thought two specific groups were targeted.

Gomez said he feels like two groups of players are targeted for drug tests: Older players and Latin players. Gomez, 32, said he and Rays teammates Sergio Romo (35) and Denard Span (34) have been tested frequently this season.

If true, this would be an explosive allegation, both because of the profiling involved, and because the testing is supposed to be random. The Joint Agreement says over and over again the tests are supposed to be random. It says so here, on page 15:

And page 16:

And further down on page 16, regarding blood tests:

So the question is whether there’s any evidence to confirm Gomez’s accusations. There were 8,281 PED tests last year across the majors, with 13 positives. The Joint Agreement makes clear that these tests are kept confidential, and so there’s no way for us to get a look at what players are being tested and when. But we can look at the lists of those players who have been suspended. You can see a couple of those lists here.

Based on this data, of players who were on major-league rosters at the time of their suspensions, 37 of 57 (65%) were Dominican, Venezuelan, or otherwise of Latin descent. That’s a stark contrast to the percentage of Latin players in the majors as a whole, which generally checks in at about 30%.

But this is just one small sample. So let’s look at minor leaguers, too. Of the 72 positive tests that are a matter of public record, 44 (61%) are Hispanic or of Latino descent. That’s awfully close to the same percentage as we see in the major leagues overall. Sources with the percentage of Latinos across all baseball are more scarce, but estimates generally put the figure at about 42%. Once again, we see the percentage of suspensions for Latin players is well higher than the overall percentage of Latin players.

Now, these numbers are based on publicly available data which is admittedly incomplete. However, troublingly, others have reproduced these findings. The excellent Tom Verducci found so last year:

Over the 2016 and ’17 seasons, of the 49 minor league players who tested positive for steroids, 65% were pitchers from the Dominican Republic—and 20 of those 32 pitchers tested positive for Stanozolol, or Winstrol, the infamous drug of choice for Canadian sprinter Ben Johnson way back in 1988.

But this is not simple. Remember that we’re just using positive tests as a proxy for tests themselves, which is a very imperfect method. For instance, Latin players could — for whatever reason — simply be more likely to use performance enhancing drugs. There is evidence of that as Verducci notes; in the Dominican Summer League, use rates have reached as high as 3% of players. Nor is this a new problem: as early as 2008, the Law and Business Review of the Americas was citing statistics showing that as many as 11% of Latin players were testing positive for PEDs. That article eventually concluded that “baseball players from Latin American countries are using steroids at a disproportionately high rate.”

Moreover, that Latin players are testing positive doesn’t mean they’re the only ones with a drug problem. As Verducci notes,

[o]ver the 2016 and ’17 seasons, of the 64 minor league players who violated the drug program for drugs of abuse, amphetamines or other banned stimulants, 90% of them were American-born players who came out of high schools, junior colleges and four-year colleges.

Here’s what we can conclude from the numbers. There is little doubt MLB has a problem. What remains to be seen is what that problem is. Are more Latin players testing positive because they’re targeted more, or using more, or both? On the one hand, to the extent MLB is targeting specific groups, it may be doing so simply because those groups are disproportionately more likely to use PEDs. On the other hand, however, that isn’t actually what the Joint Agreement permits. After all, the Agreement does specify the tests have to be random, so race-based targeting, particularly without cause, would almost certainly be illegal under federal law. At this point, we don’t have enough information to say with any certainty what the answer is, besides that this merits more investigation. That’s really the point of all of this — sometimes we have much less certainty than we’d like to think.