Gonzalez, Kemp, Bonifacio, Bourn, and Young
What do these fellow batsmen have in common?
• Adrian Gonzalez
• Matt Kemp
• Emilio Bonifacio
• Michael Bourn
• Michael Young
Well, probably a lot, seeing as how they all share a profession, but today let us examine a particularly unique distinction: The fact that they collectively represent the top five BABIPs of the 2011 MLB season.
Let’s find out how much was luck and how much was repeatable.
For this exercise, we will be toying around with FI wOBA (Fielding Independent wOBA) and slash12’s xBABIP.
Before we go any further, it is important to clarify that BABIP — batting average on balls in play — is not purely luck. I am just as guilty as most when it comes to wrongly convicting a hitter or pitcher of Luck Crimes when their BABIP strays from .300, but the truth is a lot of factors play into a player’s BABIP — from changing skill levels to improved defensive alignments (see: Carlos Pena) and so on.
FI wOBA tries to step in the direction of better divining what is luck and what is skill in smaller, single-season-sized BABIP samples. With the five top BABIPs in the league, we have an interesting divergence of inputs which makes for a useful case study.
Here is a look at each player’s data:
Key:
wOBA — Weighted on base average.
FI wOBA — Fielding Independent wOBA.
CaB FI wOBA — FI wOBA using career BABIPs.
BABIP — Batting average on balls in play for 2011.
slash12’s xBABIP — A BABIP regressor based on batted ball types.
Observations:
For each of these five league leaders, their CaB FI wOBA was lower than their FI wOBA. That means that each player would regress offensively if they returned to their career norm BABIPs in 2012 — which is beyond obvious. With FI wOBA, though, we can offer accurate, numerical predictions.
For instance, Adrian Gonzalez, despite having his career-best wOBA in 2011, would actually have had a wOBA beneath his career average (.375) if his BABIP had been at normal levels. Why? Because his walk rate went down as did his home run rate. Nearly the same is true for Michael Young, who had one of his best offensive seasons despite depreciation in other areas of his game, namely his home-run rate and walk rate.
Matt Kemp, on the other hand, had a CaB FI wOBA equal to his real-life wOBA. In fact, his FI wOBA says he should have hit 18 points better than his already gaudy 2012 wOBA. Why? Consider this: He had career highs in home run rate, walk rate, and stolen base rate. His BABIP may have been high in 2011, but his peripherals were just as high.
Using slash12’s xBABIP, we can peer into a more clear picture of future expecations. When a hitter has a BABIP higher than his career norm, it does not automatically imply he is lucky. Perhaps he is hitting more line drives (which can lead to fewer home runs, but more total offense) or more ground balls (which have higher BABIPs than fly balls).
Here is where it gets interesting: Given the preponderance of ground balls and line drives from Michael Young and Emilio Bonifacio, we should expect them to both have higher BABIPs than they actually did in 2011 (thus the green cells in the above chart). This means their hitting profiles suggested wOBAs higher than they their real-world 2011 wOBAs.
Bonifacio (24%) and Bourns (26.6%) had career-high line drive rates in 2011, while the other three had their second-best or third-best career line drive rates. Combined with the other facets of their game, only Matt Kemp would do worse (though still laughably good) if they all maintained those batted ball rates.
Of course, it would be silly to expect consistency in the realm of batted ball rates — they fluctuate wildly (not to mention the possible reporting errors in publicly available data), and compared to walk rates, strikeout rates, and the other inputs to FI wOBA, batted ball rates are simply hard to trust.
Ultimate Conclusions:
Adrian Gonzalez tattooed the ball in 2011. He may not repeat his line drive rate in 2012, but given his still-strong home run per fly ball rate, a decrease in LD% would likely just result in more homers. In other words: His BABIP may go down, but his wOBA should stay near .400, all else equal.
Matt Kemp is insane.
Emilio Bonifacio can repeat his surprisingly successful 2011 season — it’s within possibility — but unless he is suddenly seeing the ball better, resulting in more line drives and fewer fly balls, he will need to change his approach (more walks, fewer Ks is the ideal solution).
Michael Bourn has more troubling signs around him. If he cannot sustain his crazy line drive rate (and that seems unlikely), then his offense will fall back on the haunches of a near-career-low walk rate, a high strikeout rate, and a disappearing home run per fly ball rate. Without BABIP luck in 2011, Bourn would have been just a pair of legs without a base to steal.
Michael Young has the ability to hit a 22% to 25% line drive rate, and unsurprisingly has a .338 career BABIP. He probably will never repeat nor improve upon his 2011 BABIP and line drive rate, but his offense still should be good — though still reliant on the whimsy of batted balls.
But these days, what isn’t?
I’m not sure why you draw different conclusions on Bourn and Bonifacio. They have almost the exact same career batted ball profiles, plate discipline figures and HR/FB rates. They both have career BABIPs around .340, in both cases in-line with their respective xBABIPs. They almost identical offensive performances in 2011. If 2012 sees a return to their career norms they should both perform a bit worse than they did in 2011, each batting around .270-.280 rather than .290-.300. In light of all of this, I don’t see why the outlook for Bonifacio should be “good” while the outlook for Bourn is “bad”.
Actually, I have another issue with your analysis: Adrian Gonzalez’ 2011 line drive rate of 21.2% was right in line with his career figure of 21%. It was his GB and FB rates that deviated from his prior performance. His 2011 BABIP was not at all supported by his batted ball profile like Bonifacio’s and Bourn’s were. All three should regress, but not at all for the same reason (a decrease in LD%) in all cases.
As a side note, if Gonzalez’ batted ball profile returns to career norms we should expect about 33 home runs over 700 PA (assuming he continues his streak of playing 159+ games). But again, this is not due to a change in his LD%, but a change in the mix of FB and GB.
It’s not stated directly in the article itself, but Bonifacio’s walk rate was up and his strikeout rate was down last year, indicative of a better approach at the plate. (Anecdotally, Bonifacio always struggled with high fastballs in the past, but got a lot better with them in ’11. He looked like a different hitter last year — a smarter one.)
Whereas (and this is stated directly) Bourn’s plate discipline was worse last year than in previous years — he was just bailed out by a bunch of line drives.
Applying a flattening measure like career stats is going to overlook trajectories — it’s going to take into account a lot of stuff about Bonifacio that does not appear to be true any more, and it’s going to camouflage some worrying signs that have emerged lately in Bourn’s hitting.
Great article! As someone who knows most stats in the Fangraphs database but not lots others such as FI wOBA and xBABIP that aren’t listed in the player cards, thanks for explaining them thoroughly.
I always like more ways of clarifying which players’ BABIPs were significantly luckier than others.
Bonifacio is interesting and his speed is tantalizing. His LD% increased 5% from 2009 to 2011 which corresponded with a jump in BABIP from .312 to .372 during the span. As you asked, is Bonifacio suddenly seeing the ball better resulting in more line drives? I don’t know, but looking at his plate discipline numbers he has clearly evolved as a hitter from ’09 to ’11. His O-Swing% is only slightly better, but the big difference is that he is swinging 8% less at pitches in the zone. It annoys me that the Plate Discipline stats (which I love) don’t include pitches seen / PA, but from the numbers we do have it looks like Emilio is being more selective. Thus his spike in LD% may be sustainable, giving some confidence that he could post another .280+ 35+ SB season. The concern is that an early BABIP dip could send him to Ozzie’s bench, so it’s a risk.
I would imagine Fenway inflates BABIP because of the way it is built, that might have at least something to do with AGon. He hits the ball opposite field a lot which results in what used to be a HR or fly out into a double off of the wall.
This is certainly possible. Though I suspect only the best of the best of hitters (see “Boggs, Wade”) can “use” Fenway to their advantage, time may (emphasis on may – I am not predicting this, please do not flame me) show A-Gon to be in that category. His 2011 home/road splits certainly fail to contradict the hypothesis: his road BABIP of .358, while still unsustainably high and not supported by his batted ball profile, was well below his home BABIP of .402. As I said, only time will tell if Gonzalez can take advantage of Fenway to inflate BABIP, but we can’t yet rule it out.
The wonkiest of the wonky. Love it.
“I am just as guilty as most when it comes to wrongly convicting a hitter or pitcher of Luck Crimes when their BABIP strays from .300”
“Luck Crimes” seems like a missed/under-realized opportunity. Crimes Against Normativity? Meanslaughter? I dunno, spend some time with it.
Good article.
I lol’d at “meanslaughter.”
“Matt Kemp is insane.”
I can’t wait for his attempt at 50/50 😉
So now let me get this straight…according to this and a couple of previous posts this past spring, BABIP isn’t about luck after all. That’s now the official wisdom about BABIP. You know, as a fan of this site – not a stat head or a regular poster, just a plain old consumer (and yes, you’re free of charge. Mega props for that!) – I have to say when you guys vomit up a major waffle like this your credibility takes a bigger hit than you realize. You’ve spent literally years telling us that BABIP is all about luck.
The other day for the first time ever I heard a Red Sox radio announcer actually use BABIP to explain variations in a player’s performance. He carefully explained that it was “an advanced statistic” that attempts to measure how luck affects both pitchers and hitters. I thought “Hey! Score one for the good guys!”
As a Commish of a couple of Fantasy leagues. I have to drag my fellow owners kicking and screaming into the 21st century when it comes to “advanced stats.” So these prevarications are not just annoying, they make ordinary baseball fans (the vast majority of the people I hang with and go to ball games with) even more suspicious of your whole enterprise. And that sucks. Get it together!
Part of the wonder of advanced metrics is that our understanding of them continues to grow. It would be damaging to the cause if we were to learn something that contradicted an early conclusion and simply dismissed it because people were finally buying into what we had been saying before. Part of being intellectually honest is continuing that deeper examination of both new and current advanced metrics and correcting ourselves if need be.
It’s refreshing that the sabermetrics community can redefine and adapt so quickly, think about how long it takes the mainstream coverage to correct horrible stats which we all know are worthless. Baseball stats have been in the dark ages for decades, there’s no reason to bemoan the fact that as we push forward a few corrections need to be made.
Beware of people who claim to have THE ANSWER about anything. Anyone who doesn’t believe that knowledge and understanding evolve with time and new discovery is only selling you religion. I suppose religious faith in BABIP is better than religious faith in AVG, but it’s still allowing yourself to be stifled intellectually. Also be sure that you don’t misinterpret “this is our best guess to date” as somehow meaning “this must be how it is”
Give ’em hell, Rotorooter. Absolutely love it!!!! I can hardly sit still at the anticipation of waiting for the usefulness of FIP to die off.
How very negative of you.
FIP is built on two principles:
I) K-rate, BB-rate, HR/FB rates correlate well year-to-year (so these components are relatively skilful; that is, a lot of signal to not much noise);
II) a linear combination of these factors does a good job of predicting ERA (so it measures well to the “real world”); and
That’s it. I really don’t understand why people malign FIP. It’s like maligning OBP because it counts all hits the same, or maligning OPS for counting four walks equal to a home run.
I’d like to know if LD%, GB% and FB% for hitters correlate well year-to-year. If they do, and better than K, BB and HR rates do, then there’s no excuse for not using SIERA instead of FIP. But if they correlate worse, there’s totally an argument for not using them.
There’s a spectrum here of being inclusive on one side and being indicative of skill on the other.
Should read “if batted ball types for pitchers”, etc.
In any case, I think we can agree that context-neutralization is an important step in determining a player’s value. It’s the reason RBIs are a bad stat. And normalising BABIP is sometimes part of that process.
Arguing over calling it “luck” is a waste of time. The argument over how much is “luck” is not. One of you guys should go figure it out.
Nobody ever said BABIP for hitters was “about luck” — I mean, there may be fluctuations from year to year, and those are luck, but I don’t think anyone ever claimed that batters couldn’t affect their BABIP by various skills (being able to hit to specific places, speed, etc.)
The claim was always that PITCHERS didn’t have (much) control over where the ball was hit to. Completely different thing. And it’s still true to an extent, it’s just that the question of how much is “much” is called into question.
What ECN said.
If someone was saying that BABIP was purely luck, then it wasn’t here on FanGraphs. Ever. Seriously: Ever.
RotoRooter,
Most of the people reading this are committed to the validity of statistical analysis as a tool for understanding baseball. You really don’t have to worry that a couple of articles on this blog are going to undermine the legitimacy of sabermetrics as a whole.
Besides, I don’t think there’s suddenly been a sea change in the Official Fangraphs Opinion about BABIP. I think most writers on the topic acknowledge that “luck” is shorthand for “something we don’t understand which probably isn’t repeatable.” And people have said for “literally years” that batters have more control over their BABIP than pitchers do. I agree that people have begun to tease out the factors that go into BABIP more recently, but that’s not a bad thing. If you aspire to produce objective research with a scientific[ish] methodology, you should accept that new evidence may force you to change your conclusions.
“Besides, I don’t think there’s suddenly been a sea change in the Official Fangraphs Opinion about BABIP.”
You’re so far off on that statement. Perhaps you haven’t been on this site this entire year. If you don’t think Steve Slowinski’s “Jeremy Hellickson and Re-Defining BABIP” article represented some sort of seismic change in Fangraphs’ opinion about BABIP, please point me to the articles where anything of this sort had been published before it. Because I can provide literally hundreds of links to old articles claiming luck and nothing else was driving BABIP(something that pretty obviously was never true to the majority of people who watch baseball).
Again — hitter BABIP, or pitcher BABIP? They’re completely different things. If you can show me anyone who’s been claiming that hitter BABIP was dependent solely/almost solely on luck… well, let’s see you find someone like that, and then we’ll talk. 🙂
This conversation has already devolved into people responding to the overall thrust of the post. 1) You say something bad about fangraphs. 2) People like fangraphs. 3) It is the natural reaction of people to blindly defend things they like. 4) Thus people will respond to the tone rather than the content of your post in order to defend fangraphs.
Doesn’t matter that you’re right
The reason I’ve always dismissed BABIP is that it doesn’t (to me) take into account that hitters aren’t, nor the pitches they see, static entities. Players get better sometimes, change their swing or approach, pitchers change their approach to them. There are just too many variables to give BABIP any relevance IMO.
I’m not a stat geek, but I’d love to see someone analyze Howie Kendrick’s development. My theory is, totally based on my eyes, is that (A) because he can put the bat on the ball so well, and (B) because he’s not selective enough at the plate, he therefore hits into more outs than he should. I really think if he waited for a “better” strike, (A) his hits (and walks to a degree) would rise significantly.
I don’t understand what it means to dismiss BABIP. It is just a number, measured from what actually happened on the field. You don’t need to know anything about statistics or sabermetrics to calculate it. Unlike many other advanced metrics it makes no adjustments for anything. It is a record of events just like BA, OBP, ERA, etc.
I think he means he dismisses it as a predictive tool? Which seems a little drastic to me — just because you recognize a measure’s limitations in predicting future performance doesn’t mean you have to start ignoring it.
On an unrelated note, Kendrick is my least favorite player in the league right now.
Matt Kemp is insane, pretty simplistic conclusion for a writer on fangraphs.
well he is
Bonifacio, hitting behind Reyes and before the power guys in that lineup, should be seeing a lot of good pitches to hit. It’s unlikely he sees an increase in walks.
Is it possible that Young intentionally sacrificed power for more consistent contact last year? Career bests in K% and SwStr% too.
Hey Brad, when are you writing that Cubs book “Looking for Mordecai?” And have you seen the awesome ad campaign for MLB 12 The Show? Their commercials feature two recurring dreams of mine (and I’m sure yours), Kate Upton and the Cubs winning the World Series. They even have my favorite hot dog place, The Wieners Circle, in the ad.
I love what you’re doing with xBABIP (why can’t it be available on Fangraphs?) and FIO. Keep up the great work