Barreled Balls and Preventative Pitching

Last week on this website, I wrote about how the Red Sox’s offense stacked up in its ability to get the barrel of the bat to the ball. So-called Barrels are difficult to identify, but the people in the Statcast lab have done work to get us there. By looking at the exit velocity and launch angle of high-value batted balls – those that typically go for home runs or extra-bases, having at least a .500 batting average and 1.500 slugging percentage – they determined a range of launch angle-exit velocity combinations that represent barreled-up balls. Batters want to hit the ball hard, but also with a certain angle off the bat so as to maximize the likelihood of damage. In Statcast terms, this means hitting the ball with an exit velocity of at least 98 mph and at a launch angle between 26 and 30 degrees (expanding as the exit velocity increases). As I noted last week, I recommend you check out this article and the MLB.com glossary entry for barreled ball to see a visual of the idea. To this point my focus has been on what the batter wants, but there are pitchers involved in any barrel and they need considering. Pitchers want the opposite of what the batter wants: they want to avoid having their pitches result in a dent in the side of a brick building.

As they did for batters, the great team in the Statcast lab has created leaderboards that allow us to find the pitchers who are the best at limiting batters from barreling up their pitches. These can be defined per batted-ball event (Brls/BBE) or per plate appearance (Brls/PA). So we can look at how often pitchers allow barreled-up balls when their pitches get turned into balls-in-play, but also factor-in the fact that some pitchers allow fewer balls-in-play than others due to things like strikeouts and walks. With our minds now re-oriented to the pitching side of the batter-pitcher matchup, let’s look at how pitchers fared in limiting barrels-against.

Before getting right into things I think it is worth considering that, now more than ever, starting pitching and relief pitching are two very different roles. As such, I evaluated the roles separately. Starters were defined as pitchers who started more than half their games. This is a crude definition, but it provides a decent distinction. With the starter/reliever categorization made, I then limited the data to only include starters who had at least 60 batted ball events, and relievers who had at least 45 batted ball events. 60 BBE was used as the threshold for the starters because it matched the threshold I used for batters last week, and 45 BBE was used for relievers because it was reasonably above the dataset minimum of 30 to provide a strong sample. These thresholds were mostly arbitrarily selected, and therefore can be adjusted if someone thinks there is a reason to do so. In any case, using these BBE thresholds leaves 533 pitchers (216 starters, 317 relievers) in the sample.

For the 216 starters, the average Brl/BBE was 6.96 percent (SD = 2.35). As was the case when looking at the batters, barrels do not happen very often. The starter who was the best at limiting opponents from barreling up his pitches: Alec Asher of the Phillies (1.2%). I was not sure I knew who he was until I looked into this metric, but it turns out he had five nice starts for the Phillies last year. My knowledge of 2016 starting pitchers aside, the list of the top-25 guys at limiting barrels includes many of the pitchers you would expect to be at the top of a list like this: Noah Syndergaard, Gerrit Cole, Jose Fernandez, Kyle Hendricks, Aaron Nola, Rich Hill and Jake Arrieta. Changing things to be on a per-PA basis lowers the average rate to 4.36 percent (SD = 1.61) and flips the top spots, moving Brandon McCarthy (0.6%) into first and Asher to second (0.9%).

How about the Red Sox starters? As a group they had the third best ERA in the American League last year, and this is reflected in the group’s ability to avoid opponents’ barrels. Other than Clay Buchholz and Drew Pomeranz, all of the 2016 Red Sox starters who met the BBE threshold were better than the league average. Note that being better than average means having a low rate. Steven Wright and his dancing knuckler was a standout for the Sox last year; he was ninth best in Brls/BBE and 11^th in Brls/PA. Oh, and just as you expected, Sean O’Sullivan was a top-30 pitcher in terms of barrel-limiting. He will certainly be missed. Fortunately the Red Sox acquired a reasonable replacement, Chris Sale, who was not far behind in limiting barrels-against.

Henry Owens did not have 60 BBE last year, but I included him because this provides another bit of evidence that he is not ready to pitch in the major leagues. He has serious control issues (career 4.66 BB/9) and when he allows contact, it is hard contact more often than anyone would like. Given this, and his struggles this Spring, it might be time to move on from the Henry-Owens-is-a-reasonable-depth-option thought process. The rest of the list – O’Sullivan notwithstanding – shakes out mostly as expected. I am interested in seeing Pomeranz’s San Diego/Boston splits for this metric. From what I recall, and this is likely unfairly biased, he gave up all kinds of rockets during his time with the Red Sox last year; he certainly has this Spring, which doesn’t exactly instill confidence.

On to the relievers. For the 317 relievers included in the sample the average Brl/BBE was 6.3 percent (2.99); relievers allow barrels slightly less often than starters. Joba Chamberlain did not allow a single barrel among his 49 batted ball events. He was one of three relievers to do so last year, along with German Marquez (Colorado) and Xavier Cedeno (Tampa Bay). Accordingly, putting things on a per PA basis does not change the top spots, but as is typical when making this adjustment the average barrel rate drops: 3.71 percent (SD = 1.86). As for the Red Sox’s relievers, much like the starters, they were for the most part better than league average.

Craig Kimbrel is an interesting example of the importance of considering the barrel measure on both a per batted-ball and per PA basis. Per batted ball, Kimbrel was slightly worse than league average, meaning that when opponents managed to put the ball in play he was not much different from the average reliever in having that contact come off the opponents’ barrel. But when you consider how often he strikes out (and walks) batters, his barrel rate is much more palatable. Noe Ramirez did not make the BBE threshold, but I included him here for similar reasons as with Henry Owens: he should probably not be considered a viable major league option. If the 2017 Red Sox have Ramirez and Owens throwing many innings then I suspect something has gone wrong.

As I noted when looking at barrels from the batting perspective, this is supposed to be a process metric. We are interested in finding pitchers who, due to their pitch repertoire, pitch movement, deception, or other factors, are able to consistently avoid having opponents barrel-up their pitches. For batters I suggested this is about swing and approach consistently resulting in contact on the barrel. If it is a process metric we should expect some year-to-year consistency. This was the case for batters, as the year-to-year (2015-to-2016) correlation for barrel rates was quite high (0.747 for Brls/BBE, 0.742 for Brls/PA). This turns out not to be the case for pitchers. The Brls/BBE correlation across 2016 and 2015 for pitchers who met the BBE threshold criteria by pitching role in each year was only 0.13; it was only slightly better for Brls/PA: 0.19. Pitchers changing roles between seasons and survivorship bias (those who frequently give up hard contact are not kept around) likely have an effect here, but this is still interesting. For those interested, here are scatterplots of the batters and pitchers year-to-year Brls/BBE and Brls/PA. Taken together, these correlations suggest that it is the same batters who are barreling the ball year-to-year, but they are not necessarily doing it off of the same pitchers each year.

While the relation between years may not be as clear for pitchers, here are the changes for Red Sox pitchers (grouped by role; negative numbers mean the rate was lower (better) in 2016):

The emergence of Steven Wright is clear, as it is for Fernando Abad. Wait, that seems odd. Perhaps, as for Drew Pomeranz, it would be interesting to see Abad’s team splits for these barrels measures. Otherwise most of the Red Sox pitchers stayed fairly similar across the two seasons.

In the end, the majority of the 2016 Red Sox pitching staff allowed barrels at better rate than league average. This certainly helped them post a top-10 team ERA. The extent to which it is replicable in the coming season is unclear, as the available data suggest the year-to-year barrel-against rates involve a lot of noise. Regardless, barrels-against are not the whole story, but rather just one component, and the additions of Chris Sale and Tyler Thornburg, and improved defense in left field and at third base should keep the run prevention side of the 2017 Red Sox in the top third of the league.

Photo by Kim Clement – USA TODAY Sports