During last night’s game, as the Braves seemingly took called strike after called strike and generally fared ineffectually against Steven Matz, I began to wonder whether this was an approach issue, in terms of a silver bullet situation. The Braves have the league’s second-highest walk rate and came into last night’s game with a bottom-10 overall swing rate, with diminished rates of both swinging at strikes and (especially) chasing balls. Matz, meanwhile, has not been a stellar major leaguer — he had a dominant 2016 but struggled with injuries in 2017 and ineffectiveness last year — but if he has a calling card, it’s that he throws a boatload of strikes. Last year, only four starters with 70 or more innings hit the zone more often than Matz, who did so 49 percent of the time. In 2017, he was 27th out of 178 pitchers with as many innings as him or more. In his great 2016, he was second only to Rich Hill. You can see the thinking: combine guy wot throws strikes with team wot is content to wait out pitchers, and a generally sound approach for 2019 can turn sour in a hurry.
Now, I was more than happy to run with this as an explanation for last night, in which Matz dealt the Braves their second-toughest opposing start by FIP, xFIP, and Game Score (v2) so far. But, this is as good of a time to reinforce this lesson as any: supposition is fine, and a great starting point for exploration, but let the data speak. In this case, here’s what the data said:
Red shading in a column reflects things that are “worse” for the Braves and better for the pitcher. To save you the trouble of sorting or counting, basically:
- Yes, Matz filled up the zone again. But so did Pivetta, Freeland, and Marquez, each of whom the Braves handled or crushed.
- The Braves weren’t actually waiting out his strikes and getting burned. Their z-swing rate was relatively high compared to other starters. While they’ve gotten burned by low z-swing rates before (Nola, Arrieta), this wasn’t the case.
- They even chased relatively little (third-lowest of the season), but in the end, that didn’t help.
In the end, it wasn’t an approach thing, but just one of those baseball things that happens over the course of the season. Matz actually failed to miss many bats, but the Braves weren’t squaring things up when they connected (.279 wOBA, .252 xwOBA against Matz — once again, the second-worst team result of the year against a starter by both measures). It was very much a “launch angle fluctuates and didn’t fluctuate awesomely” game for the Braves — of their 14 struck balls off Matz, half had negative launch angles, and two were hit above 35 degrees, which is mostly can-of-corn land. Only two were struck at the “ideal” 15 to 25 degree angle for contact (higher angles are good with more exit velocity, too), and both of those were run-scoring hits (triple, homer, hit at 20 and 19 degrees, respectively). It happens, onto the next.
There’s also probably some mention in order regarding the shoddy strike zone calls last night. The way I think about strike zone umpiring, other than “Get thee to a place/time where robot umpires exist, posthaste!” is whether it’s possible to draw a non-stupid shape around the differentiated strikes and balls being called. For Braves’ pitchers last night, the zone wasn’t that stupid:
Yeah, there’s a few misses, but you can draw kind of a lopsided oval and still separate the orange from the blue. For the Mets, though... yeesh.
This was mostly a failure, and is yet another in a long line of obvious examples of the need to move towards something more accurate than a trained eye and the human element for officiating the rules of the game. There’s no non-stupid shape to be drawn there.
Adding injury to insult was the fact that beside the Mets getting net four calls while the Braves got net zero (both teams had two strikes called balls; the Braves also had two balls called strikes while the Mets had six balls called strikes), two of those blown calls came on full counts and would have been Braves walks had they been called correctly. While neither happened in super-duper-high leverage and there’s a greater chance they wouldn’t have changed the game’s outcome than them having a definitive impact on the standings, this sort of occurrence is as particularly annoying as it is feasibly preventable. Johan Camargo’s eight-inning strikeout looking was on the most badly-missed call of the night, which is just gross.
So far this season, of borderline pitches taken in the zone (n = 185), umpires have ruled 39 of them (over 21 percent) to be balls. Of borderline pitches taken outside the zone (n = 249), umpires have ruled 64 of them (over 25 percent) to be strikes. Rather than pace of play, shifts, or a bunch of other stuff, this lack of clarity and consistency with regard to the basic building block of baseball is probably something MLB should look to address sooner, rather than later. Although they are testing automating aspects of the strike zone in the Atlantic League this year, MLB should ask itself what rate of missed borderline calls is acceptable, set some public benchmarks for it, and actually start taking steps to move closer to that rate if they’re not going to implement an automated system with all deliberate speed.