An Attempt to Explain Yasiel Puig’s BABIP
Been talkin’ about BABIP lately. Let’s talk about BABIP again. Let’s talk about Yasiel Puig, and his BABIP.
Last week, I wrote a post on Starling Marte, in which I examined his extraordinarily high batting average on balls in play. I had a hypothesis, and that hypothesis was confirmed. It was far from revolutionary. I knew that Marte was fast, and then I found out that he hits a bunch of line drives and never hits pop-ups. Then I also found out that those three things alone can explain more than 50% of the variance in a player’s BABIP. Again, that’s really nothing new.
The metric I created, BIP Score, featured Marte prominently near the top. Also near the top were a whole bunch of guys with BABIP’s above .330. Yasiel Puig is another guy with a BABIP above .330. It’s way above .330. During his time in the MLB, only two qualified batters have a higher BABIP than Puig. But he’s nowhere to be found in the top half of the BIP Score leaderboard. From the post:
Not everyone with a high BABIP scores well in BIP Score. Yasiel Puig, for example, owns a career .366 BABIP — higher than Marte’s — but actually has a negative BIP Score, thanks to his low line drive and average pop-up rate.
I felt like that warranted an examination of its own. This post is that examination.
I guess, first, we’ll take a look at that BIP Score. That’s how this all started anyway. To get BIP Score, I simply summed the z-scores of every qualified batter’s line drive rate, infield fly rate and speed score and scaled it so 0 was league average. It’s admittedly a quick-and-dirty metric, but the higher the BIP Score, the more likely it is that a player should be able to sustain a high BABIP.
With this methodology, Puig clocked in with a BIP Score of -0.3. To get a sense of the context, let’s look at the other guys around Puig who also clocked in at -0.3.
| Name | LD% | IFFB% | Spd | zLD% | zIFFB% | zSpd | BABIP | BIP Score |
|---|---|---|---|---|---|---|---|---|
| Brandon Moss | 20% | 9% | 3.8 | -0.7 | -0.1 | -0.1 | 0.291 | -0.3 |
| David Ortiz | 20% | 7% | 2.3 | -0.5 | 0.6 | -1.0 | 0.289 | -0.3 |
| Yasiel Puig | 17% | 9% | 5.6 | -1.9 | 0.0 | 1.0 | 0.366 | -0.3 |
| Desmond Jennings | 17% | 11% | 5.9 | -1.6 | -0.5 | 1.2 | 0.295 | -0.3 |
| Justin Morneau | 22% | 9% | 1.8 | 0.3 | 0.0 | -1.3 | 0.310 | -0.3 |
| Domonic Brown | 20% | 12% | 4.3 | -0.5 | -0.8 | 0.2 | 0.278 | -0.3 |
| Pedro Alvarez | 19% | 7% | 3.5 | -1.1 | 0.4 | -0.3 | 0.277 | -0.3 |
So, you see why it makes sense to take a look at this. In Puig’s group, the closest BABIP to his is 56 points lower, and it’s from the guy who plays in Colorado. Of course, all this really means is, Puig has been slightly below average thus far in his career at a combination of hitting line drives, avoiding the pop-up and being fast. But shouldn’t that mean his BABIP should be… y’know, not .366?
The first thing I wanted to look at that isn’t included in BIP Score was batted ball direction. Going up the middle is the best place to hit. The ability to go up the middle or spray the ball to all fields is an indicator of someone with a “pure” swing, someone who can’t be shifted. Shifts kill BABIP. Think Joey Votto. Think Joe Mauer. I expected Puig to demonstrate this, to some extent. I was wrong.
- PULL
- League: 40%
- Puig: 41%
- CENTER
- League: 35%
- Puig: 35%
- OPPO
- League: 25%
- Puig: 24%
Puig hasn’t yet demonstrated the ability to use the whole field any more than the league average, so we can cross that one off.
The next thing I wanted to look at goes hand-in-hand with speed, and it turns out it’s something Puig does quite a bit:
Probably not the best example, but really I just wanted to make a .gif of Yasiel Puig beating out an infield hit because damnit it’s February and I miss baseball. But, point is, Puig does this a lot. Since he entered the league, Puig’s got 42 infield hits. That gives him an infield hit rate of 12%. Last year, Billy Hamilton was at 11%. So was Dee Gordon. So was Jose Altuve. Part of it is that Puig’s fast. He’s not as fast as those guys, but he’s fast. The other part is that he hits a ton of ground balls.
Since he hasn’t hit any line drives, those extra balls in play have to go somewhere, and more often than not, they’ve gone on the ground. Puig’s ground ball rate since he entered the league is the same as guys like Denard Span and Jon Jay, and he’s done a better job than those guys of beating them out.
Grounders, paired with speed, are often a good thing. It’s weird, though, Puig hitting so many grounders. For someone with his strength, it seems wrong to have his name mentioned around guys like Gordon, Altuve, Span and Jay. Feels like we should want him to hit less grounders. Feels like we should want him to hit more line drives and elevate the ball. Then again, it’s worked, and a low line drive rate isn’t necessarily a bad thing.
Qualified hitters, over the last two years, with the lowest line drive rates: Hunter Pence, Mark Trumbo, Yasiel Puig, Josh Donaldson, Nelson Cruz, Jose Bautista. All good hitters, those guys! Really good hitters. In fact, it gives us a pretty perfect comp for Puig, in someone that’s gotten by as a very good hitter with the same unique skill set. These are career numbers:
| Name | LD% | GB% | FB% | IFFB% | IFH% | Spd | HR/FB% | BABIP |
|---|---|---|---|---|---|---|---|---|
| Hunter Pence | 16% | 51% | 33% | 11% | 10% | 5.1 | 15% | .319 |
| Yasiel Puig | 17% | 51% | 32% | 9% | 12% | 5.6 | 15% | .366 |
By all accounts, it’s essentially identical, and this should be almost everything that could reasonably explain one’s BABIP — luck excluded. Puig pops out a bit less than Pence, and is a bit faster, but Pence also goes up the middle more often than Puig (not pictured). Yet, in the end, we still have nearly a 50-point difference in BABIP.
So what gives? My gut tells me Puig can’t keep this up. Of course, I’m not exactly going out on a limb by saying the guy with the .366 BABIP can’t keep it up. But with a guy like Marte, you see a strong line drive rate, you see an avoidance of pop-ups, you see the speed, and you see a .350+ BABIP that actually seems pretty reasonable. With Puig, all you really see is the speed. He doesn’t hit line drives, he doesn’t avoid the pop-up, he doesn’t spray the ball. All he’s really got going for him, in terms of his BABIP profile, is the speed that leads to infield hits.
Of course, this isn’t to say that Yasiel Puig is going to fall off a cliff. This isn’t even to say that Yasiel Puig isn’t going to continue being a great hitter — he probably is. After all, Pence has still sustained a .320 career BABIP with essentially the exact same profile, and Puig’s strength and athleticism make him unlike almost any other player in baseball. But there’s still something that gives me pause, in that it’s pretty tough to explain Yasiel Puig’s BABIP.
August used to cover the Indians for MLB and ohio.com, but now he's here and thinks writing these in the third person is weird. So you can reach me on Twitter @AugustFG_ or e-mail at august.fagerstrom@fangraphs.com.
As a dodger fan this bummed me out. Something that i’ve kind of noticed just by watching him all the time is that the grounders he hits are a lot of times hit in unanticipated places. What I mean is that I’ve seen countless times him being able to pull a ball a foot off the plate and slice it right through the 5 6 hole, particularly breaking balls. I think his strength / deception in standing relatively far off the plate causes teams to leave the 5-6 hole open and yet pitch to him in spots that for normal hitters wouldn’t lead to a hard grounder in the 5-6 hole. Just something i think I’ve noticed; no idea if its really accurate.
If your anecdotal observation means anything, it helps support Fagerstrom’s thesis. A hitter who is badly fooled on a breaking ball, pulls off the ball, but still manages to find the hole between short and third, is not engaging in a repeatable skill. It’s dumb luck. And it will always be dumb luck even if you’ve observed it “countless times,” which we know is a vast overstatement since Puig has had all of 940 Big League at bats. No hitter intends to pull off the ball. If pulling off the ball is by design, it is a design that will end your career early.
Yes you are probably right. But I also think his repeatable skill sets are far from plateauing. I think its very possibly he continues to improve his plate discipline with experience. His adjustment from his first year to second year was pretty encouraging.
As a casual watcher of the Dodgers, I’ve noticed that Puig hits the hardest ground balls I’ve ever seen. A lot of the time, he just crushes them. I’d hypothesize that his grounders find more holes than usual, and infielders are getting to balls in unfamiliar positions that make is difficult or impossible to get a good throw on the ball. Couple that with Puig’s speed, and you have the makings for a high BABIP number off of ground balls.
I totally agree. I watch a ton of Dodger games and Puig hits harder grounders than just about anyone I’ve ever seen. He hits missile after missile through the 5-6 hole.
What these guys said. Its amazing what you can learn by watching players play baseball.
I figured the same, but without exit velo data to back it up, not much to do there. Can’t wait until that becomes a thing
Is there any hope on the horizon that it will become a thing in the near future? Is there a reason why pitch fx is publicly available but hit fx isn’t?
Could be mistaken, but I think StatCast should make this a reality.
I remember seeing some GB, velocity data that suggested that ~ 90 MPH, there was a ~60% chance a GB would go for a hit. If my recollection is accurate, that jibes with the speculation about Puig.
I was about to make the same comment but decided to read all comments first. His grounders are hit as hard as anyone in the league, infielders have no time to react. He will always hit for average regardless of the strikeouts and popups
Is it possible that Puig’s Babip is explainable by hit speed and angle? Puig hits the ball really hard so, even if he hits grounders at a league average rate, fewer defenders will be able to reach his balls and he will post an inflated babip. Two years of a great hitter posting above average babips is a reasonably good indication that he does something differently. I would imagine that Pence’s Babip is a floor and Puig’s rookie season babip(.373 I think) is a ceiling. I’ve been drafting for a repeat of last year with slightly more power.
Craig Goldstein proposed an interesting thought to me on Twitter. Going under the assumption that Puig hits GB harder than the average person (which would lend itself to a higher GB BABIP than most already), that might cause infielders to play back — something they wouldn’t do for the likes of Span, Jay, etc. When playing back, little dribblers like the one in the .gif above become more likely to go for hits.
No idea how much of an impact that might make, but it’s interesting to think about.
Can you track Inside Edge data to hitters? i.e., which hitters caused the highest relative percentage of difficult plays? Might be another way to get at xBABIP.
It definitely happens that fielders play back for hard hitters. Just look at where the defenders are positioned in the .gif. 3bman is two steps behind the bag, and the ss is waaay in the hole. Normally for a player with good speed, you’d play in, but Puig’s hard hitting makes it necessary to play back.
In that sense, then, Puig’s BABIP is a function of _two_ variables: grounder velocity and foot speed. Either one alone might be defensed, but the combination gives him a ‘bigger infield’ to play on, in effect. That wouldn’t suggest that he’ll improve his present rates, but gives some support to the idea he’ll maintain an above average number of infield hits from his present approach. I think most pitchers will be more than happy to trade infield grounders for LDs, so I wouldn’t expect much change in the pitching plan against him so long as Puig keeps expanding the zone to contact marginal pitches.
Infield playing back gives defenders more time to get to the ball but also creates larger holes and he does hit the ball harder than average. I’m not sure it is a good strategy to play a deep infield on Puig. You will not get him on slow rollers and you will need to make a great play on hard hit balls.
https://www.youtube.com/watch?v=F8KWZiBwnd0
http://m.mlb.com/sf/video/topic/39049312/v36736841/stllad-gm1-puig-extends-the-lead-with-an-rbi-single
Thanks for doing this! I have been wondering about Puig’s batted ball profile and his performance. I think it might limit his ceiling to what he is doing now (which is WOW) but he can only go down as he slows and runs into luck dragons. His projections represent the high end of what we might see. Harper, by contrast, hasn’t done nearly what Puig has, but he hits liners and flies with great authority, suggesting he has more room to grow. Because of their batted ball profiles, I’m taking the under on Puig’s projections, and the over on Harper’s.
Agreed, and for exactly those reasons.
For Puig to over his projections he needs to cut down on his o-zone swings, increase his LD%, or both. I’m doubtful he can do the latter; the former is a matter of motivation, and might happen some years on in his career. Otherwise, the Puig we’ll get is the Puig we’ve got.
I think some of it is that his BABIP has been slowly regressing to a more realistic number. So the .383 BABIP in 2013 is clearly not sustainable based on the data shown in this article. The .356 from last year almost certainly isn’t, either, but it is starting to become more reasonable.
The .366 career number is partly skewed by the 2013 number. I would say that he probably continues to regress to .330/.340, which will ding his slash line a bit, unless something else changes.
If the power increases or if he starts hitting more line drives or flyballs, or if he continues to cut down on IFFB, the BABIP might actually stay high or it might come down but with not much decrease in his overall slash line.
As a Puig owner in a dynasty league, my fear is that he actually gets UNlucky one of these years, posts a .275 BABIP and a slash line around .260/.330/.400. When he is relying so much on luck or hitting the ball hard on the ground (or both), an unlucky year is going to be a disaster.
You don’t normally expect a guy like Puig, who has posted a .366 on over 1000 plate appearances, to go full .275. That imagine would require the serious erosion of some tools in addition to bad luck. .300 to me seems like an unlucky season for current Puig.
First…favorite name of any FG write
Second, I still don’t buy into Marte on the basis of repeatable BABIP alone…anything above .360 is difficult to sustain for anyone, and given his poor plate discipline and contact skills, BABIP regression will hit his output more so.
While I think you’re right on Puig likely seeing a BABIP regression, he showed other improvements from 2013 to 2014 that if continued will likely outweigh the BABIP regression — increased BB% 2pts, decreased K% 3pts, cut O-Swing% 9pts, increased Contact% 8pts, and cut SwingStr% by 5pts. Given his age, its entirely likely that his growth continues.
At 2 years older, Marte’s peripherals are fairly steady, other than a 1.5pt increase in BB% and don’t project that same adjustment.
I took the unpopular anti-Marte stance in the last article but like all of these great fangraphs articles, it made me re-think. My problem is the reliance on the speed factor since that always seems the luckiest of the three components. However, after more poking around, it seems that IFH% is rising every year and those that thrive on it are just going to keep posting better BABIP.
I think Tony Blengino’s numbers would help a lot in answering this question. Hunter Pence seems to set a floor of .319 or so, but Puig is a bit faster and pops up a bit less. I assume this means that, as long as Puig keeps his speed and pop up rate, his floor would be more like .325 or .330.
What’s Puig’s BABIP on fly balls? If that, too, is elevated, I’d like to know Blengino’s numbers about speed off the bat. I don’t know if that would fully explain a BABIP of .356 or even .366, but it would suggest his true talent BABIP is even higher than .325 or .330.
There may also be a batted ball classification issue at play here as well. Puig may hit a large number of short liners that bounce toward the far reaches of the infield and get classified as grounders. C’est possible.
Perhaps this is a thing, but I think it’s more likely that a LD gets misclassified as a FB than a GB, and the high GB% would support a low LD%
And we already know that IFFB% is unreliable. I forget what FG writer was discussing how team proprietary data apparently disagrees significantly, in some instances, with what the public data classifies as an IFFB.
Agree with Yo. Maybe he hits more griners than most. Maybe he is defended in a way that permits infield hits (The 3rd baseman looks really deep in the gif). Maybe the defenses he faces most are unusually bad at converting grounders to outs. Maybe there is some other reason. Great article to get the brain a’thinkin.
This is what BP’s Craig Goldstein proposed on Twitter. That Puig hits grounders harder than your typical high-GB guy, causing defenses to play deeper than normal, leading to more infield hits on dribblers like the one above. An interesting theory.
I enjoyed the article. I think you more or less answered this with the comparison to Pence. Sure their BABIPs are still 47 points apart, but an extra 1% LD for Puig, extra 2% IFH, 2% less IFFB, and moderately superior speed, I think those add up to something tangible. Maybe that’s the difference between 25 or 30 points? I’m not sure, but at that point they’re close enough that a few hits changes it.
Puig has gone on cold streaks, where his BABIP drops to normal levels, and I keep hoping that he is establishing a new, lower normal for his performance. But then he’ll whip off another HOF-level hot streak fueled by an insane BABIP, and my hopes are dashed again.
.376 / .442 / .598 with a .475 BIP, 258 PAs, June 3 2013 – Aug 11 2013
.245 / .321 / .417 with a .282 BIP, 244 PAs, Aug 11 2013 – Apr 13 2014
.373 / .463 / .677 with a .427 BIP, 189 PAs, Apr 15 2014 – June 1 2014
.238 / .317 / .341 with a .296 BIP, 142 PAs, June 2 2014 – Jul 8 2014
.429 / .500 / .786 with a .523 BIP, 64 Pas, July 9 2014 – Aug 3 2014
.238 / .330 / .333 with a .309 BIP, 215 Pas, Aug 4 2014 – Oct 7 2014
Keeping my fingers crossed that last July was the last BABIP-explosion hot streak of his career.
If you dig down to individual streaks, that may not be any more variation than normal. Since we know luck and defense play a role, the smaller a sample you use, the greater potential effects those two factors will have, and on top of that you are purposely selecting for the times when that may be the case. I would be interested if that variation is greater for Puig than for other players, but I kind of doubt it.
What you say is of course wise and numerically informed.
I have however seen many guys come up from the minors (or over from Japan or Cuba), and run wild for their first couple months. Eventually, opposing teams notice that the new guy is a force to be reckoned with, and they start to look at video tapes, spray charts, Pitch FX kind of stuff, and the like. They then make adjustments in pitch selection and defensive positioning. A flash-in-the-pan can’t counter-adjust. A truly big league hitter can, and does.
Clearly, Puig wouldn’t keep up the 1.040 OPS/.475 BABIP of his first ten weeks in the bigs. How much of a drop off he would suffer from other teams adjusting to him was the question for me.
In aggregate, since his first “hey, here I am” 258 PAs in 2013, he’s had another 873 PAs in 2013 and 2014, playoffs included. His line has been .285 / .369 / .468, with a .346 BABIP. If we give him another 10 or 15 points drop off in BABIP, that looks about right to me for a true talent level.
Interesting that you say that, because that further dropped level is right about what I expected from Puig coming into the league, along with the high in-season variance between hot and cold streaks he has already demonstrated. Puig hits the ball hard, and at present has good straight line speed. He swings at way too many non-strikes to sustain high levels of productivity without far greater contact skills than he has historically demonstrated though. Puig may continue to tighten his swing rates; then on the other hand, he may loose a little speed by the time he does.
. . . Frankly, I’m amazed Yasiel’s numbers have held up even as well as they have so far. It’s taken crazy hot streaks like that July to do it. Somehow, I don’t think those are going to be as sustainable going forward. If Puig isn’t going to really hurt pitchers with the HR with his big swings, he’s boing to be fed a steady diet of pop-up and GB pitches. Hard to continue to be a star from those alone.
The specter in this room is obviously batted ball authority. We’ve long wondered if some pitcher BABIP outliers just generate weaker contact, and now Puig provides us with a case whose BABIP is unexplainable, except perhaps for the possibility that he makes harder contact on the balls he hits.
Totally agree. I think the “exit velo” statistics will be the next big thing to consider in player (hitter) analysis. I could see batted ball profiles becoming more predictive than the actual result of the play, which will always have some kind of luck component.
And just to add to many of the comments above, anecdotally I’ve also noticed that Puig hits harder grounders than just about anybody.
Southern California baseball fan and Puig keeper league owner here.
1. Puig’s adjustments from year one to year two made me happy.
2. His power dry spells worry me (they always smell like a pulled muscle, I think he’s constantly hiding some nagging injury and when he gets healthy he goes on his hot streaks)
3. Saying his BABIP is attributable to hard grounders might be the case for about 50% of the time, but the other 50% is due to the following.
Time after time I see him make contact in the most awkward way possible. He puts the ball in play like no other, but he’s hitting over the top or makes the weakest contact even though he’s swinging like Thor. The ball comes off his bat weakly or at a weird angle and it’s rarely in the same spot. The swing speed and his size fool the fielder time and time again.
I think one of two things will happen in the future.
1. He will continue to do this and defenses will adjust and play in or at normal depth and his BABIP and AVG will go down.
2. He will get even more selective at the plate (I think he already has a good eye, and has shown he can make adjustments) which will result in a higher walk rate and harder contact increasing his doubles and HRs.
I hope it’s the later, but I’m worried that the pressure to make contact in the 3 or 4 spot will result in outcome number 1. I also wonder that one of the reasons he put the ball in play so often was when he was hitting in the 2 spot with Dee Gordon on base. Which then makes me wonder….
Was Dee Gordon being on base a contributing factor to Puig’s high BABIP?
Another thing: I feel like speed score misrepresents Puig’s true speed. He has always had issues being caught on stolen base attempts, which would drag down his SPD, but in reality that’s because of poor technique and not poor speed.
This is huge. SPD definitely accounts for SBs and SB %, both of which Puig has not mastered. That has little to do with his actual, raw speed. I’d imagine his explosive speed from the box is among the best in baseball.
First, August, excellent article again.
Second, it was great reading the above as commenters together appeared to be unraveling a mystery. I definitely sign on to the theory that Puig hits hard ground balls and that the defense has to play him a little further back. Whether on reputation or otherwise, a guy with Puig’s speed is going to get more INF than Gordon or other light hitters.
Finally, as to that speed: http://m.mlb.com/video/?content_id=33676413&topic_id=6479266 . Puig gets over 21 MPH in the outfield in this clip. The guy is just amazing.
I think even if that BABIP comes down, which even Puig-fans agree it will, that’ll be nicely offset with the continual growth of his other skills. He’s only 24. What a special player.
Really thoughtful stuff here, both the article and the comments.