He's not just the drummer for thoroughly unsettling black metal whirlwind Krallice; Lev Weinstein is a long-suffering Mets fan with much more patience than some of us who already never want to see Travis d'Arnaud step into a batter's box again. For those of you waving your dick around/huddled in a corner about your team's first month and a half, here's a rational and well-reasoned introduction to the wild world of "sample size" and "batting average of balls in play."
Listen, it's only fucking May. I understand the impulse to gripe and to cast ill-portending auguries into August. Dude gave up seven runs yesterday! Motherfucker is on pace to have the worst batting average in Major League history! I get it. I also understand the early-season giddiness when your team doesn't suck and players are mashing. At this rate, he'll hit 59 homers! Yeah, no.
I get it. I do it, too. You're--we're--being fucking stupid.
Even more than a month into the season, all those numbers and projections--and the accompanying excitement or hand-wringing—are utterly meaningless.
Baseball is weird. The game is weird in ways that go beyond the treasured platitudes about squarely hitting a round ball with a round bat. That said, the impossibility and constant failure that such a statement implies are at the heart of this weirdness. It’s a game in which the best who have ever played still fail at their jobs far more than they succeed.
The fact that you did your job spectacularly well one day really is no guarantee that you'll continue to do so. It's not even an indication that you're particularly good at your job. It's a long fucking haul, and one day of going 4-4 with two doubles and a home run doesn't mean a whole hell of a lot if you finish September with a .250/.300/.350 slash line. The hot start that stud of a prospect is having? Fucking chimerical. But by the same universal laws, that atrocious slump your superstar du jour is mired in is just as likely to be illusory.
For sure, skills decline as one ages, and sometimes those prospects are Mike Trout, but people usually don’t fall off cliff faces so much as inexorably roll down the hill, and the only Mike Trout to come along in a loooong while is… Mike Trout.
This all boils down to fundamental issues of sample size. Out of 16 regular season contests, the production an NFL player displays in one game is a rather large and potentially revealing indicator of what kind of player he is, and what one might expect him to do in games 2-16. One MLB game out of 162 tells you jack shit. Really, one month out of a season tells you jack shit.
Let's put it like this: A player will probably have somewhere around 700 plate appearances in one season. Up to mid-May, they've probably been up to bat maybe 100+ times.
If you're like most of us idiots and you're sure that a well-regarded prospect is a bust because his average is floating in Mendoza territory and he looks like a double play machine (looking at you, Andrew), take a deep fucking breath. If he’s actually allowed to demonstrate his capabilities at the major league level for something approaching a season, the hapless dude will have hundreds more opportunities to get on base, etc. Is the guy walking? Is he going deep into counts and showing patience and discipline at the plate? Does his swing look un-palsied and compact? These are probably all better indicators of what this guy’s numbers will actually look like than what one can glean by looking at his inchoate early season stats.
At some point in the course of a season, however, numbers start to actually tell you something. We all know this at an intuitive level. We know that it’s hardly revelatory for a guy to be batting .415 on April 6, but if he’s sporting that average on September 6, it’s fucking earth-shaking.
Thanks to people far smarter than you or I, we can get way more specific than that intuitive analysis would ever allow. We can pinpoint the exact milestones at which the numbers we use to evaluate player performance actually become useful metrics and not just more horseshit fodder for pundits craving talking points on which to wax philosophical.
A couple years ago, Fangraphs posted this extremely useful guide. This thing is pretty great. It's your one-stop-shop for seeing how many plate appearances, batters faced, balls in play, etc. a Major Leaguer needs before his stat line becomes an accurate predictor of performance. In other words, how many of those 162 games does a guy need to have played in before you're justified in thinking he's fucking terrible? Looky here.
Some things really stand out: It takes 910 plate appearances--probably more than one season's worth--before a player's batting average stabilizes. That prospect hitting .150 through the first month and change? Don't fucking sweat it. That is, don’t fucking sweat it if he’s not hanging up a golden sombrero every night, taking balls down Broadway and hacking at pitches five feet off the plate with a bat that looks like it’s moving through air that’s thicker than the stuff you and I walk around in.
Those strikeouts are one of the few things it's worth the air to talk about this early on. The rate at which a guy is likely to whiff is actually one of the few metrics that stabilizes rather quickly. K rates and walk rates both in the batter's box and on the mound resolve around this point in a season, and that makes sense. Those are the kind of skills that other people on the field can fuck with less. Once the ball's in play, defense and dumb fucking luck are, too.
Let’s revisit our hapless prospect from above. If he’s hitting .150 and his OBP is .150, he probably sucks. But what if the guy is maintaining an OBP solidly above his torporous average, and he's not a strikeout machine? Chances are he doesn’t suck so much as his BABIP does. For the ignorant among you, BABIP is an acronym for "batting average of balls in play." When you make contact, how often is that resulting in a base hit? In general, both pitchers and hitters have less control over this number than you'd think. Batters manifest more control over BABIP than pitchers, but even so, deviation from the mean is rare.
It’s true that some guys tend to fucking wallop the ball, and consistently average higher BABIPs than others. But if a guy has a BABIP in, say, the high .400s, he's essentially been ridiculously lucky. There's a very, very, very good chance that a lot of those balls that have been finding gaps are gonna start finding gloves. When that happens, BABIP will drop to something more closely resembling league average, floating somewhere around .300. When a guy's BABIP drops, so too will all his other glossy early-season numbers.
This is all equally true for the player who's looking terrible early on. If a guy has a BABIP well below .300, you can be pretty damn confident that the man's luck will change. Odds are good not 100 percent of his contact will result in slow dribblers to the shortstop.
Oh, what's that? I don't understand? Every time this guy makes contact, the ball screams off his bat? Well, sure. Line drives are gonna fall for base hits more than fly balls or grounders, but most of 'em still get caught. Let's talk about whether your guy's inflated BABIP is the real deal after another 820 plate appearances when the stat actually stabilizes.
Listen, I'm not suggesting that you're always going to be wrong about that special something you see in that one guy, or the complete lack of something you see in the other. I'm merely suggesting that you're almost always going to be wrong. You're not a scout. You're absolutely without a doubt going to fuck it up if you're going off of what you've observed in the minute sample size that is a baseball season in its spring. And you're abso-fucking-lutely beyond the even the most far-flung fractional possibility of a doubt going to fuck it up if you're going off of the stat line generated by the minute sample size that is a baseball season in its spring.
The idiot who races ahead of everyone in the first 10 minutes of the marathon is not on fucking pace to obliterate world records. Relax, guy.
- Aug 28, 2014
- Aug 25, 2014
- Jul 23, 2014
- Jun 13, 2014
- Jun 13, 2014
- Jun 13, 2014
- Jun 10, 2014
- Jun 9, 2014
- Jun 9, 2014
- Jun 4, 2014