An Open Letter to ESPN about BABIP
Dear poor intern who works at ESPN pumping out preseason player comments:
I know it’s hard. You have to make some kind of bold prediction for every single player, with pretty much squat to go on other than what fans can already get from Fangraphs. You’re not a scout. You haven’t seen these players swing live lately, maybe never at all. And even if you had, you’re not a hitting coach. You wouldn’t be able to tell what has gone terribly wrong, or what a guy finally figured out last year that has changed everything. Most of the time your best bet is to take a page from Marcel the Monkey, and predict that a player’s numbers are going to be the same as his career averages, slightly adjusted for whether he’s on the still-learning or declining side of his career. But you got this job because of your reputation as a statistical whiz, and that’s just not going to cut it. Your boss wants some kind of numerological wizardry, for you to predict the future with your big brain and the lastest SABR science. I get it. I’ve been there.
But please stop using BABIP every time you get stuck. You’re hurting statistics, shredding whatever scraps of credibility have been won over the last decade, turning back the tide of superior metrics creeping onto scoreboards and into newspapers and even coming out of announcers mouths, reinforcing the stereotype of sabermetricians being pimple-faced teenages who know nothing about how the game is actually played, who don’t even watch it, preferring instead to read boxscores and crunch numbers long after the games are over from the safety of their parents basements, blah, blah, blah-dee-blah blah blah…
I know, I know…years ago Voros McCracken turned stats on their head with BABIP. He started a huge controversy and was eventually hired by the Red Sox after stumbling over the fact that, in a nutshell, pitchers don’t have much control over anything other than strikeouts and walks. Once contact has been made, batted balls fall in at the same rate no matter who is on the mound (after defense, park effect, etc, is taken into account). Therefore, pitching to contact is not what it’s cracked up to be, and if a pitcher has a really low or really high BABIP, you can say that over that period he has been “lucky” or “unlucky”, because nobody maintains a very high or very low BABIP; eventually every pitcher’s will end up the same. (Almost…for example, we now know that closers can have lower BABIP’s, but not by all that much). This makes for a great tool to gauge the amount of statistical fog when dealing with small sample sizes for pitchers, and an incredible tool for pumping out quick fantasy baseball comments.
But it just doesn’t work the same way for hitters — and nobody (even Voros!) has ever maintained that it does. It would be like saying that a spike or fall in batting average has to be luck, because for hitters, BABIP is just the same damn thing with K’s and HR’s removed from the equation (really two of the last things you want to exclude when trying to decode if a streak or slump or bad year is due to something tangible). While it is true that if a player has an extremely uncharacteristic BABIP over a period of time, those results should be taken with a grain of salt, you could just as easily say that about batting average because they’re practically the same damn stat. Obviously, nobody would call every .200 hitter a victim of luck — all that’s really going on here is that if you don’t hit somewhere in the realm of league average, you’re not going to get major league at-bats for long.
Still, BABIP is regularly used these days as a diagnostic when there is a suspicion of someone’s skills having deteriorated — and in that sense, it just doesn’t tell you anything at all. It is of zero value for making future predictions. It is nothing more than a huge misunderstanding, a misapplication of a SABR truth that has gone viral. I repeat, there is absolutely no universal number that all batter BABIP’s (or averages) bounce randomly around, as there is for pitchers.
Because a very low BABIP could mean player is simply making lousy contact, which unlike pitchers, batters are certainly capable of affecting. Compare the BABIP of a great contact hitter like Joe Mauer (note: his exceedingly high BABIP is not because he’s beating out infield hits with his speed as per Fangraphs explanation of Ichiro’s consistently high numbers) over the last three years — .348, .373, .343 — to a noodler like John Mcdonald — 260, .269, .235. That’s some 100 points of consistent difference, and obviously not luck-based. So when a player’s BABIP plummets 100 points, it could be a fluke — or it could mean he used to make contact like Ichiro and now, for whatever reason, he’s making contact like John Mcdonald. You have to look deeper to have any idea what’s really going on.
Last year, BABIP was everywhere in the discussion about David Ortiz’s early struggles, which were obviously the result of mechanical difficulties and not luck. This year, batter BABIP is continuing to bloom — I’ve seen at least a dozen snippets like this one about Aaron Hill (Carlos Quentin is another good one).
Think you’re unlucky? Even though Hill hit 26 homers in 2010, his batting average plummeted to .205 thanks to a .196 batting average on balls in play. To put that in perspective, that was almost 100 points below his career mark and 30 points below the next-lowest mark in the majors. That’s unlucky. Also consider that Hill’s strikeout rate held steady and that his walk rate improved, and it jibes with the scouting that says Hill hit the ball hard right at a lot of people for much of the year.
I’d like a word or two with some of these scouts, because anyone who watched the Jays last year knows that Hill wasn’t getting unlucky in the slightest. His contact SUCKED. His whole season was one weak pop-up, head down, jog to first after the other. Hill looked like a completely different hitter, his beautiful compact line drive swing gone long and loopy – and a little deeper delving into the numbers agrees: his fly ball rate soared from his terrific 2009 (41.0), past his career average (41.4), to insane heights (54.2). His line drive rate also fell from 2009 (19.6) past his career average (18.5) to untold depths (10.6). His infield fly percentage (one of my favorites because it’s a 100% guaranteed out – like a hidden strikeout, but still included as a ball in play) rose from 11.6 to 12.9. Spraying balls like this will clearly and provably lead to a consistently much lower BABIP. And so the numbers actually overwhelmingly agree with what the old-timers would say: Hill’s swing went to hell last year. Luck had nothing to do with it.
Now yes, players tend to return to their normal numbers. That’s usually a good thing to point out because it’s really rather amazing how even the most terrible results over extended periods of time can be due to random fluctuation. We can’t help but think we see what looks like conclusive results of a player’s skill level having changed that are in fact just baseball doing its crazy and unpredictable thing. (An example from the classic “The Book” is that even over 1000 batters faced, one in every 20 pitchers will have an ERA more than a run over their true skill level over that time). So it is still a very good bet that Hill (or any other major leaguer with a highly uncharacteristic BABIP) is going to bounce back.
But that is not because of any statistical techniques. It’s because nobody with that kind of pedigree and years of success hits under .200 in his prime for long unless something has gone terribly, terribly, wrong physically (it was his hamstring, FYI), something which he is likely to be able to recover from. But when you state authoritatively to the millions and millions of people who read your comments that a comeback is predicted by pure numbers with no need for a turnaround in his on-field performance or condition, an entire range of astute fans will begin to think that all stats are stupid and all analysis useless because they don’t jibe with most obvious truth they know and can see and hear about baseball: that a sweet swing and solid contact from a good hitter makes a ball jump off the bat, while a poor hack from a guy lost at the plate sends up a dying quail.