Is Goldy’s BABIP Sustainable?

You know the story: in 2012, Paul Goldschmidt was quite good. In 2013, he made The Leap into the MVP discussion on the back of a 36 HR, .302/.401/.551 season, finishing with a wRC+ of 156. He was almost exactly as great in a Frieri-shortened 155 wRC+ campaign last year, and since coming back this spring, he’s taken it to a completely different level: 189 wRC+, a bigger jump from his previous two seasons than when he had made The Leap from 2012’s 124 wRC+ line.

League batting average on balls in play tends to be right around the .300 mark. Last year, it was .299; this year, perhaps in part because of increased shifting, it’s down to .296. When a player — hitter or pitcher — sees a BABIP much off of that mark, we tend to expect regression. Some fast players can sustain higher levels by beating out more hits; some seem to stay above that level because they hit the ball hard more regularly. Few manage to stray from that mark regularly by more than 20 points. Goldy’s BABIPs from 2012-2015: .340, .343, .368, .386. Yes, Goldy’s BABIP this year is 90 points off of league average. It’s enough to make a smart man wonder.

Specifically, this one:

@thepoolshot what is the highest, sustainable BABIP? Goldy’s is high, but he’s had high-ish his career, and it’s increased.

— OJ Carrasco (@youknow_OJ) June 8, 2015

Among qualified hitters so far this year, Goldy ranks just fifth in BABIP, behind a few fast players (Dee Gordon .421, DJ LeMahieu .411, Anthony Gose, .395) and Kris Bryant (.391), who may turn out to be a near-clone of Goldy at some point soon. Right behind them are Jorge Soler (.383) and Nelson Cruz (.382), two guys with a reputation for hitting the ball hard.

We’re late enough in the season that we can’t give luck all of the credit for a BABIP this high, and yet we’re still early enough that luck has a lot to do with it. It’s mostly earned, and yet there are probably some hitters just as deserving of hits at that rate who haven’t had the luck to match. This is the perfect time for this question: there’s something here in Goldy’s .386 BABIP, and yet just 60 games into the season, it’s not its own evidence.

The Quick Answer

Bad news for Gordon and LeMahieu: those BABIPs probably aren’t going to happen. The highest BABIP recorded by a player qualified for the batting title since the D-backs came into existence is .404, by Jose Hernandez in 2002. Manny Ramirez is the only other player to record a BABIP over .400 in that span. Last year, the highest BABIP was that of Lorenzo Cain (.380). The year before, three players did better than that: Chris Johnson (.394), Joe Mauer (.383), and Michael Cuddyer (.382).

The mere fact that I just mentioned Chris Johnson probably makes you suspect what it makes me suspect: even a year’s worth of stats contains some luck. When I combine the 2013 and 2014 seasons, Goldy ranks 8th with a .353 BABIP; Johnson still adorns the top of the list at .369, but the next three players, Starling Marte (.368), Yasiel Puig (.366), and Mike Trout (.363) all pass the luck smell test, as all three have power and foot speed. And if those numbers are “right,” Goldy’s .353 also looks “right” — he’s no Prince Fielder, but it’s taking nothing away from Goldy to say he may not be beating out as many infield hits as Trout.

.353 for Goldy seems like a pretty good benchmark. It’s worth noting that two very good projection systems still doubted even that level; Steamer projected a BABIP of just .330, while ZiPS had him at .348. I’m still more comfortable saying .353, since his career level is .351 and projection systems tend to be skeptical. Expecting something in that range for the rest of the season seems reasonable, although he did manage to beat that by a fair margin last year and in light of his doing so again so far this year, I don’t want to be dismissive of the potential for sustainable improvement. That’s why we’ll continue to dig.

Someone Else’s Not-So-Quick Answer

Quick for us, though. According to calculations that Alex Chamberlain published at RotoGraphs just before I finished writing this today (using Jeff Zimmerman‘s formula, I believe, which Jeff Wiser has explained elsewhere), Goldy’s expected BABIP using component stats from this season is .352. Basically, the exact same answer as .353. All righty then.

The Not-As-Quick-But-Still-Dirty Answer

The Zimmerman xBABIP formula uses a few component stats for things we know affect BABIP: line drive percentage (mostly a function of launch angle); Hard% from Inside Edge (hit it hard, good things happen); Speed Score, also from Inside Edge (more speed can = more infield hits); and a coefficient for spray, since a high pull percentage means being especially shift-able. I’m not saying I can do better than that, because I can’t.

But now that we have raw hard-hit data in the form of batted ball velocity, we can do better. It doesn’t need to just be a function of what percentage of hits were hit hard according to a stringer; we can get more granular. I recently put league-wide data into four buckets to figure out what was wrong with Mike Napoli; there’s no reason we can’t use it to help figure out what’s right with Goldy.

Paul Goldschmidt hits the ball hard, pretty often. Among hitters with at least 50 tracked balls, he ranks 12th in average batted ball velo (93.17 mph — from Baseball Savant). That’s pretty badass. But that’s not even necessarily the whole story; the average of 85 and 50 is the same as 90 and 45, but chances are both of the latter ones are outs, and the 5 mph on the top makes a bigger difference in terms of whether we should expect a hit.

We could get more granular, but for this, here’s the chart from the Napoli piece for MLB averages:

The 75-89 mph bucket probably scales up significantly — 88 mph may fall in for a hit a lot more than 78 mph. But maybe not. It’s remarkable that there was really no difference in this sample between buckets 1 and 2. And yet, as noted, so much in there could affect a player’s overall average.

Out of 104 tracked batted balls, Goldy’s look like this:

In other words, Goldy has done a bit better than we would have expected, across the board. In particular, he has many more batted balls over 100 mph than the average player; take 14 batted balls from the top category and dump them somewhere in the bottom, and even if the rates held up he’d have maybe 7 fewer hits. isn’t quite BABIP, because it includes home runs (which are not “in play”) — and yet it’s pretty clear that Goldy’s hard hit percentage didn’t quite capture the extent to which batted ball velocity justifies his BABIP.

Add Goldy’s fairly good speed and high marks in terms of being shift-proof, and you have everything in place to justify a not-quite-very-best BABIP. The only missing piece: Goldy’s 23% line drive percentage is a bit lagging in terms of the current leaders this year. Considering he’s likely to keep up a strikeout percentage a bit north of 20%, a BABIP that is 30 points lower for the rest of the season probably “drags down” his average from the current .338 to something like .310 (finishing the year somewhere in between). He may be especially and unsustainably hot. But not by much.

Actually Answering OJ’s Question This Time

With everything working in the right direction — batted ball velo, launch angle, spraying the ball, and a ton of speed — I think the highest sustainable BABIP is probably something in the .360-.365 range. Goldy doesn’t have the absolute most elite line drive percentage, and doesn’t have Trout/Marte type speed — I would say someone like that could probably only maintain .350. By the end of this season, we may have cause to move that up a smidge, to .360 — but I don’t see any player with a very similar combination of skills. He’s a demo of one, which is why this question was so interesting in the first place — we may never be able to isolate luck completely, but Goldy may end up doing more than just teaching us who he is — he may teach us something about the nature of the sport, too.