What went wrong for the Braves in their Game 4 loss

If you clicked on this article thinking, “Ugh, why?” I want you to know that really, I agree with you. From the Braves’ perspective, the 2018 season has ended. It’s over, it’s finished, it’s done. Lessons learned, post-hoc analysis, and everything else, well, they’re nearly irrelevant. By the time anything gained from a look back at this game can be put into practice, there will be a new season, a new team, a new roster, and hopefully some new approaches to certain aspects of the game. The Braves had a charmed, immensely gratifying season that puts many Cinderellas to shame, to pointing to the one game that ended it all and saying, “See, but if only they had done X!” kind of misses the point.

So, to that end, I’m not going to sit here and pretend that there really is much of a purpose to the below. Instead, it’s just a series of observations that I thought could be interesting to share. The main throughline here, though, is that as tempting as it is to say the Braves lost because of Thing X, Thing Y, or Thing Z, they largely just lost because of baseball. Not because baseball was specifically wild or wacky on a Monday afternoon in Atlanta, but because it was baseball nonetheless.

I’m not going to look at the whole game. Doing so would lead to reprising a lot of well-trod ground covered in these posts for Games 1 and 2. If you don’t know by now that the time Nick Markakis hit a 73 mph slider from Rich Hill in the exact middle of the plate and turned it into a pop-up was a particularly brutal outcome, I don’t think rehashing that general idea will help. Instead, I just want to focus on the “welp, baseball” nature of three specific things.

Nick Markakis and the 3% Hit Probability Bloop

Yasiel Puig’s bloop was not really one of the bigger plays of the game. It resulted in a WPA swing of just about four percent, making it only the ninth-most-critical play of the game, and only the fifth-most-critical such play when the Dodgers were at bat. Yes, the tying run moved to third, but it was only the tying run, and there was already one out in the inning.

Nor was it something particularly mismanaged. Yasiel Puig is a bonafide reverse splits weirdo, so leaving Jonny Venters in to face him, whether intentional or not, was likely the right call in this regard.

But, any time you give up a hit on a ball with a three percent (!) hit probability, it stings. In the entire game, only two balls had lower hit probabilities. They were both generic can-of-corn flyouts. (One of those was Lucas Duda’s, which I’ll just mention here came on the exact same pitch location, pitch type, and pitch velocity from Kenta Maeda as the foul ball that was nearly a three-run homer. Go figure.) Puig’s eventual sixth-inning single was also a can-of-corn flyout, except for the “out” part.

I don’t know if anyone is going to look back and say that this ball dropped due to a communication issue between Nick Markakis and Ozzie Albies. Or that Albies should have had it. Or any other “Yeah, but!” statement. But, here’s the thing that gets me: Nick Markakis has not generally been weak at coming in to catch fly balls and liners this year.

If there’s one thing you can say specifically, it’s that Markakis has been pretty good at coming in on balls this year. Now, you could also say that Markakis has not been good at moving to his left on anything this year, and that perhaps that’s what stung the Braves here, but either way, his deficiency in this area is fairly minor for a corner outfielder. The point is just that yes, the Braves got burned on a bloop — but it wasn’t like they were doing some kind of specific thing that made them particularly vulnerable to bloops, or this specific type of bloop. It’s just a weird thing that happened.

The ball traveled 178 feet. 170 to 190 is generally the infield-outfield “nexus” as far as fielding. On the year, the Braves were the fourth-best team at causing opposing batters to underperform their xwOBA on balls hit this distance. They were middle of the pack (17th) on balls hit this distance anywhere to the right of second base. If they had a definite weakness, it was exactly where the ball ended up — they allowed the sixth-highest xwOBA overperformance on balls traveling 170 to 190 feet and in the rightmost quartile of the field. There’s nothing they could have really done about this, other than install boosters on Markakis’ cleats, but that exact confluence of events (a bloop specifically placed where the team has had trouble getting to it) was one of the pain points of the game. That’s baseball.

Charlie Culberson and the go-ahead grounder

When this back-breaking hit happened, I really wanted to curse the baseball gods for the cruel twist of fate they had delivered to the Braves in the human form of Charlie Culberson. Not that any of it is Charlie Culberson’s fault, mind you. It’s not his fault that he’s not a particularly great defender (as best we can tell, anyway). It’s not his fault that Dansby Swanson was injured and couldn’t play. It’s not his fault that the Braves chose to start him at shortstop instead of one of their other infielders. It’s not his fault that the specific plays in the sixth inning that involved him were just out of his reach. It’s also not his fault that the team never sought a better backup infield option, and it’s definitely not his fault that he went on a never-before-seen xwOBA outperformance run that cemented his roster status despite many indications it should have at least been contested, if not rejected outright.

But, fault or not, Culberson was unable to corral one grounder through the hole from Enrique Hernandez in the sixth inning, putting the tying run on base as his diving effort was woefully short. That ball had a 47 percent hit probability, and I don’t know if a different shortstop gets to it, but it also wasn’t the most sorrow-inducing grounder in the inning.

Instead, after Culberson was beat to his right to put Hernandez on base, he ended up getting beat to his left on David Freese’s go-ahead pinch-hit single. That ball, too, was not a laser up the middle, with a 54 percent hit probability and a 94 mph exit velocity. But it avoided Culberson’s glove all the same.

When it happened, the immediate knee-jerk thought was, “Dansby Swanson would have gotten it.” That statement is tempting: Culberson barely missed the ball, we suspect that Swanson is a much better defender, what else do you need? We’ll never know for sure, though. Even if we knew for certain that Swanson would have recorded the out there, so what? There wasn’t a choice between Swanson and Culberson for this game or this inning or even this series (I liked the idea of starting Swanson and not allowing him to swing, but the Braves aren’t that daring), and to the extent there was a choice, it was a choice of finding a different backup infielder in March, when there was no earthly idea that Culberson’s defensive shortcomings would allow an opponent to score the runs that would prove fatal to the Braves’ 2018 season.

All that aside, though, I wanted to dig deeper. What I wanted to show was how annoying this confluence of events was — that Freese didn’t hit anything particularly unique, that the Braves cut down balls like this all the time, that this wasn’t a problem, until it was, in this one crucial moment. (And yes, it was the most crucial. No play in the game had a higher Leverage Index or a higher WPA swing.) But, perhaps like the instance earlier in the inning, with Markakis not flagging down a stupid bloop, it wasn’t annoying per se, just baseball-tastic.

Using Baseball Savant, I was able to determine that Freese’s grounder was hit at a radial angle of around -7 degrees. To think about radial angle, basically just figure that the vector from home plate to second base is 0 degrees, the right-field foul line is 45 degrees, and the left-field foul line is -45 degrees. So, -7 degrees is mostly up the middle, but somewhat to left. What we don’t have numerically is where Culberson started, degree-wise, on the play. But what we do have are these bits of data:

Data bit #1: Charlie Culberson’s average starting position at shortstop places him at -14 degrees of radial launch angle.

Data bit #2: Dansby Swanson’s average starting position at shortstop also places him at -14 degrees of radial launch angle.

Data bit #3: One way to think about defensive effectiveness is to think about the difference between wOBA and xwOBA on specific balls in play. This basically gives you an adjusted-for-difficulty ranking of how well the team’s fielders cut off balls in the specific areas. I previously did a bunch of work on this and found that launch angles below nine degrees are generally fieldable by infielders (though balls hit with a launch angle above 7 or so and at 110 mph are basically not fieldable unless the infielder doesn’t have to move); another way to do this is to use a cutoff of 170 feet from the plate in terms of distance, but these differences are semantic as they mostly cover the same array of batted balls.

In any case, if we restrict our sample to righty hitters facing righty pitchers, launch angles of nine or below, and radial angles commensurate with the Freese hit, we get the following:

• Radial angle of -8 to -6: Braves allowed a .539 wOBA (4th-highest in MLB) on a .272 xwOBA (8th-highest in MLB) across 17 batted balls. The difference was the third-worst in MLB, i.e., the Braves generally did not position to get these specific balls. But, 17 batted balls is a pretty small sample.
• If we expand the radial angle more broadly, to -9 to -5, we get: 42 batted balls, .470 wOBA (6th-highest in MLB) on a .241 xwOBA (22nd-highest in MLB). The difference was the fourth-worst in MLB.
• Let’s expand the radial angle to get a sample of at least 100 batted balls in this area, or -11 to -3 in radial angle. With this, we get: 100 batted balls, .425 wOBA (13th in MLB), .251 xwOBA (19th in MLB), and the 10th-largest difference.

The thing is, we already know that the Braves are pretty good at infield defense: their overall xwOBA-wOBA gap for all batted balls with a launch angle of 9 degrees or lower was in the bottom 10 among MLB teams, or a top-third defensive infield contingent by this measure.

So this just makes Freese’s game-winning hit more baseball-y: the Braves have an effective infield defense, but get burned on one specific component of that infield defense that they don’t do that well (mostly because they tend to play some righty bats to pull more). Just like the Nick Markakis bloop, of course it happened that way: that’s baseball, I guess.

Nor can you even blame that Freese dagger on poor scouting or positioning, even though it got through. For one, this is Freese’s spray chart this year when he faces a righty pitcher and hits a ball with a launch angle of 9 degrees or lower with a radial angle of 0 or lower (pull side):

This spray chart shows balls that are fielded, so you can see that a few got to the outfield. The red arrow is the one that sunk the Braves on Monday night. The blue arrow points to the Braves’ standard positioning for Dansby Swanson and Charlie Culberson. Without the gift of foresight, and given this spray chart? Where would you want the shortstop to be standing when Freese is batting against a righty? With the gift of foresight, you can see how it will all go wrong.

As a side note, I tried to do something more rigorous, applying a very rudimentary topographic analysis to a radial angle-by-angle analysis of Freese’s grounders. That analysis suggested more or less the same thing that the Braves did: when Freese faces a righty, position your shortstop somewhere between -20 and -10 degrees, probably around -16 to -14ish. Again, without knowing exactly where Culberson started that play in terms of degrees, that seems like what they did during that fateful play, and it’s consistent with what they’ve done all season. You can only really blame baseball.

Manny Machado and fastballs

This has already been too long (I guess I have a hard time letting go?), so this will be quick. But it won’t necessarily be painless.

Before the series started, I noted this about Manny Machado:

Machado is a hard guy to pin down. It requires a more detailed analysis to see whether he changed his approach after being traded from Baltimore, but the surface stats suggest that he started hunting four-seamers in Los Angeles, giving up his ability to stay balanced against changeups, sliders, and even sinkers in the process.

After his trade to the Dodgers, Machado put up a .352 xwOBA. That’s a pretty scary number, but the truth is that facing a righty pitcher really neutralized him, tumbling his xwOBA down to a manageable .325. But, that’s not because Machado has issues with righty fastballs, as his xwOBA against them was actually .371 since the trade. Instead, it’s non-fastballs from righties (.241 xwOBA) that really plagued him since the trade.

As a major leaguer, Chad Sobotka has thrown his only non-fastball pitch, a slider, 36 percent of the time. In the Braves’ final game of the 2018 season, Chad Sobotka threw six straight fastballs to Manny Machado. The sixth was deposited into left field for a three-run homer. To add numerical insult to the existing scoreboard injury, Sobotka’s sixth fastball ended up right in Machado’s wheelhouse, the part of the zone where he’s had the highest xwOBA this season (middle third vertically, inner third horizontally).

In the end, it didn’t matter. The Braves didn’t score another run anyway. But you have to wonder about the thought process. The Braves put Chad Sobotka on the roster due to a thought process that could most charitably be described as “recently focused,” given that he had, to that date, pitched all of 14 and a third major league innings. However, that sort of “recently-focused” aptitude was nowhere on display — the Braves, up to that point, threw Machado a steady diet of non-fastballs, and he complied by flailing at them miserably, with an 0.048 xwOBA against them, and with a set of results comprising seven strikeouts and a pop-up on his PAs that ended with a non-fastball from a righty earlier in the series. Perhaps they were gunshy because earlier in the game, Machado pulled a Mike Foltynewicz slider into the corner to score the game’s first run, but it was the first solid contact he had had against that kind of pitch all series.

Maybe this doesn’t fit the overall theme of this post. Maybe throwing six straight fastballs to a guy who kills them isn’t “that’s baseball” but rather “that’s bad decision-making.” But when a virtual unknown, a rookie reliever who’s only faced a handful of MLB batters in his short career, throws the same exact pitch six straight times to a hitter who feasts on them, and that hitter hits a three-run homer in an elimination game off of it, well... that’s either baseball or it’s cruelty. Or it’s both; no reason the two can’t intertwine. So it goes.