clock menu more-arrow no yes

Filed under:

Reconsidering catcher's interference

New, 15 comments

The tale of one of baseball's lowliest statistics, and the case for its redemption.

Rob Carr

You're at the bar, watching your local baseball team lose its fourth straight. The inch of Pabst left in the plastic pitcher has taken on an ashen hue. Your buddy, picking through the corpse of a plate of nachos for a greasy olive slice, throws a lazy challenge at you: what are the eight ways a batter can reach first base? (There are more than eight, but your friend's hardly an expert.)

There are the easy ones: hit, walk, hit by pitch, fielder's choice, error. And the somewhat rare ones: dropped third strike, a batted ball striking a runner. Only if you're clever, and it's the first pitcher, will you remember one of the rarer varieties: catcher interference. After all, it only happens a handful of times a season, and when it does, it's almost without fanfare: a flick of a bat on leather, the stoppage of play, a minor confusion and a leisurely stroll. It's so rare, so minor, that it's hardly worth thinking about it at all. Unless you're Bob Stinson.

Stinson, nicknamed Scrap Iron, was a rarity: a Mariners expansion pick that paid off a little. His career was unspectacular: a first round draft pick in both drafts of 1966, he was traded three times before he lost his rookie eligibility. By the mid-seventies he was proving an adequate backup catcher for the Royals, hitting in the .260s with a solid walk rate and absolutely no power. That was good enough for the abysmal M's, who took him in the 25th round, and he repaid them with a couple of productive years before wearing down the way catchers do.

StinsonStinson's career, like that of most major leaguers, is admirable and unremarkable. But he was immortalized in an early Bill James Abstract, involved in of all things the 1978 AL MVP discussion. Ron Guidry, who sported a 25-3 record that year, had been touted by some for the unlikeliness of such a record. James refuted: "The argument that what Guidry did is more unusual than what Rice did is specious... it's an award for the most valuable performance, not the most unusual. The most unusual performance was turned in by Bob Stinson of Seattle, who reached base six times on catcher's interference."

That performance set the American League record, though Roberto Kelly eclipsed it soon after. Still, James betrays a mild disdain for those six times Stinson reached base, a bias that remains today. Reaching first this way has no effect on one's on-base percentage. Like the base on balls, the play is treated as a plate appearance but not an at bat. However, unlike the more familiar outcomes, it is not factored into any calculation. As far as the numbers go, it may as well have not even happened.

...

This is strange to me. It exposes in one of our most reliable and beloved statistics, OBP, a glaring dualism: a desire to measure what happened, and what should have happened. This contradiction even resides in our own conception of the statistic: the pure definition is (hits + walks + hit by pitch) divided by (at bats + walks + hit by pitch + sacrifice flies). However, in our crude human brains, we tend to think of on base percentage as something simpler: the rate at which a batter reaches at least first base without forcing anyone out.

This second definition opens the door to some unwanted guests, however, the most obnoxious of them the batter who reaches on error. Those opposed to including the ROE, which include the majority of baseball fans, make two separate arguments:

1.       Reaching on error, in its many incarnations, is impossible to award in retrospect. This would lead to modern hitters being measured by a different OBP than their ancestors.

2.       Reaching on error is not a skill, and thus batters should not be rewarded for the mistakes of their opponents.

The former is a question that plagued OBP from its inception, since early box scores omitted the cowardly walk from its records. America, and its baseball historians, persevered. And while the task for errors may be more difficult, it's in any case irrelevant to our more focused examination of catcher's interference, which Baseball Reference has compiled with admirable reliability for the past eighty years.

The latter is more troublesome to apply to our specific problem. From an ethical standpoint, it hardly seems fair to lump CI in with ROE: the latter is a gift given by the defense after the hitter has already failed, i.e., when he has hit a ball that should be converted into an out. But with catcher's interference, the batter has done nothing wrong, and in fact has barely had time to do anything at all. From this standpoint catcher's interference far more closely resembles a hit by pitch than it does an error.

...

When James wrote about Stinson in 1978, he was obviously treating Stinson's "accomplishment" as the gift of a fickle higher power, something beyond his control. This seems intuitive. But baseball thinkers felt the same way about the walk, and for longer about the hit by pitch. Both of these were eventually credited to the batter because it was proven that they were repeatable skills, and that their practitioners were deserving of statistical acclaim. Reaching on error, as Keith Woolner found in 2001 and Dan Wade followed up in 2013, holds only a low correlation from year to year, and its leaderboards are often topped with players who defy the hustling, defense-forcing narrative.

But leaving aside the debate over whether OBP should report or editorialize, let's return to catcher'sinterference. In such a random and rare event, is there any correlation from year to year?

No.

I'm going to be honest with you. This is supposed to be the part with the pretty graphs. There are no pretty graphs.

Correlation is usually pretty easy. All you do is take one year's worth of data, and compare it to the next year's worth of data for each player. If a guy had a lot of RBIs one season, and getting RBIs is something that is related to the skill of the batter, then he'll probably get a lot of RBIs the next season. Sure, there are things that foul up the data: injuries, steroids, a crippling addiction to Sid Meier's Civilization. But when you're dealing with thousands of players, the big picture comes out. You plot all these points on a graph, and the closer those dots come to making a pattern, the more likely they're related.

Here is what it looks like when we do this for catcher's interference.

Ci1

This is an awful graph. The integrity of the data is undone by the concept of whole numbers: the granularity of the data leaves us with this blocky, nuance-free stack of blue dots. The rarity of the play, and the shortage of names on the leaderboard, also increases the variance of the data. The sad fact is that the play is so rare that even the "superstars" of CI rarely earn them consistently from year to year: they just don't happen frequently enough. Rod Barajas, for example, has earned it four times: once in each of 2001, 2005, 2007 and 2011. This is unhelpful.

So if we can't, based on my statistical acumen, prove that catcher's interference does correlate from year to year, we can at least estimate how unlikely it would be that it didn't. The following numbers will be really rough, but I hope they prove a point.

There have been 1,980 cases of catcher's interference since 1964. This seems like a lot. But there have been 8,253,657 plate appearances, so dismissing all other factors, there's about a 0.024% chance of it happening in any given trip to the plate. For the sake of simplicity, this estimate glosses over a slight upward trend overall in the past fifty years.

(We're assuming that there is no batter skill involved in CI, thus distributing those 1,980 incidents independently from each other. We're also assuming no catcher skill involved, which may be harder to swallow: after all, if a catcher interferes once, isn't he more likely to do it again? But catchers distribute themselves fairly randomly against batters, with the exception of in-game and in-series repetition. And given the lack of multiple CI games (see below), it seems like catchers learn their lesson after the first error, possibly spreading the sequencing out even more than random chance.)

Let's also assume our arbitrary hitter is relatively hale and hearty, and earns 600 plate appearances in a season. Using binomial probability, we can calculate that his chances of getting one CI in a given year under such circumstances is 12.5%, and multiple CIs 0.9%. A 1992 Roberto Kelly, in this case, would have a 0.000000000006% chance of getting interfered with the eight times in 643 chances (by, as it turns out, eight different catchers from eight different pitchers) that he did that particular year.

...

Bob Stinson isn't quite the king of earning catcher's interference, nor does he even own the flukiest record regarding the stat. Both of those honors go to Pat Corrales, who earned a free pass nine times in 858 replacement-level plate appearances. Robert Gordon, in his book "Then Bowa Said to Schmidt," transcribes:

"Pat was a crafty guy. I can give you an example of how crafty. He figured out how to get on base through catcher's interference! He got on base a lot that way. Mauch loved that. Mauch loved guys who took every advantage they could, legal or otherwise."

We're left to wonder exactly how Corrales "stole" first base so much, or why he wasn't able to convey that wisdom to his teammates or his pupils later on with Philadelphia and Cleveland. Was he swinging at the glove rather than the ball? Cheating an extra couple of inches out the back of the box? Either way, only five players in the past fifty years have managed to incur catcher's interference twice in a game, including Stinson and teammate Danny Meyer. Corrales did it twice in the same season.

Muddying things even further is the fact that there are multiple methods by which the catcher can interfere. The most common, by far, is the swing of the batter hitting an anxious catcher's glove. But there is also the case of August 1, 1971. With the bases loaded in the bottom of the 11th, two outs, game tied 4-4, Willie Crawford of the Dodgers faced Reds reliever Joe Gibbon.

As Gibbon went into his windup, the runner on third, Manny Mota, broke for home. It was a poor choice. Johnny Bench leapt in front of the plate, caught the pitch, and tagged the runner out, allowing the winning run to score. It turned out that Bench was too quick; he read the play so quickly that he stepped on the plate before the pitch arrived, denying Crawford from his divineright to take a swing. The umpire correctly invoked rule 7.07, catcher's interference was called on Bench, and Mota scored the winning run.

Ichiro


Clearly, not every catcher's interference could be credited to the skill of the batter. On April 11, 2010, David Murphy reached first twice in a row off Ian Snell because of the call, and former M's catcher Adam Moore admitted after the game that he was simply positioning himself too close. (How Murphy was the only man to capitalize on this lapse of judgment is left unanswered.) But this is equally true, if not more so, by the HBP: though Don Baylor stuck out a nice elbow, there are plenty of unavoidable, and sometimes intentional, fastballs thrown straight to the flank. We still reward their pain.

I asked Mariners first base coach Andy Van Slyke if he was aware of his own place on the modern career leaderboards (sixteen, tied for third most since 1965). He said that while he didn't know exactly how high he ranked, he recognized his skill in it during his career. His remarks mirror the conclusions I've found:

"I will tell you that most of those occurrences happened with two strikes, not sure if that is a coincidence, but I am pretty sure it wasn't." [Note: it isn't. With two strikes, runners are more likely to steal, and catchers more likely to stick their gloves out further.] "I was a guy that always tried to stand as far back in the batter's box, you know, the guy that tried to eliminate the back white line in the batter's box. I wanted to get as long of a long at the pitch as I possibly could. Just one of those things I did throughout my career and this may possibly have played into the CI."

It seems clear that positioning, and hitting style (Van Slyke, like many of the career leaders, was a hitter who tended to use the whole field; late swings are another likely cause of nicked gloves) are elements of a high-CI hitter. He says, "It takes high intelligence to do this sort of thing on purpose, not saying that I always did that," and that's fine - a skill doesn't have to be 100% intentional to be repeatable. If Van Slyke's makeup as a hitter led indirectly to reaching base 16 extra times, he should get credit for that, just as we give credit to Craig Biggio for crowding the plate.

It seems like a small thing, because it is a small thing. But it's not so small to Bob Stinson; in his infamous 1978 season, he was credited with a .346 OBP, when in fact he reached first at a .356 clip. Pat Corrales in 1965 leaps twenty-one points, from .323 to .344.  They deserve those numbers. After all, they did whatever it took.