Keller was better. We take a look at why and how.
The last two starts for Mitch Keller have been quite different. A week ago, he didn’t experience much success against a pretty good San Diego Padres’ lineup; then, on Thursday afternoon, Keller pitched well, ultimately earning a no-decision in the Pirates’ 4-2 win over the Detroit Tigers at Comerica Park.
On April 15, when the Padres were in town, Keller got the start in a game Pittsburgh would ultimately lose 8-3. He pitched 3.1 innings, gave up nine hits, allowed seven runs, walked three, and struck out one batter. It wasn’t a storybook performance, but that’s been common for the young righty through his first couple years in Major League Baseball.
In 2019, Keller made 11 starts with the Pirates. Over those 11 starts, he compiled a 7.13 ERA and 164 ERA-; the silver lining that year was his FIP (3.19), which suggested Keller was suffering from a host of bad luck.
In 2020, however, those numbers swapped. Keller had a much more desirable 2.91 ERA over five starts, but saw his FIP balloon to 6.75, which means we shouldn’t have expected him to pitch as well as he was.
Now, in 2021, his ERA is right back up with his 2019 mark (7.16), while his FIP has leveled off somewhere in the middle (4.87). The latter number is where we might expect Keller to ultimately settle in his career with the Pirates. He might never become the front-of-the-rotation type pitcher everybody hoped he would, but he might slot in as the four or five guy (depending on the strength of the rotation).
To get back to his recent performance in Detroit, he pitched five innings, allowed two runs on five hits, while striking out five and managing to not walk a batter. By all indications, that’s a pretty good day. The question now is this: Was this good performance indicative of Keller’s talent or improvement from his last start? Or was it because he was pitching against the Tigers, a team which carries the fourth worst wRC+ in baseball at 83? (By comparison, the Padres are currently 15th in baseball at 93.)
Using Baseball Savant, the numbers in both of Keller’s outings were fairly similar. He threw 83 and 80 pitches, respectively; maximum exit velocities of 107.3 and 109.5; average exit velocities of 94.1 and 93.7. That story remains true for many variables in Keller’s outings, so let’s look for outliers.
Against the Tigers, Keller threw 59 percent fastballs versus 63 percent against the Padres. Of those 59 percent against Detroit, 27 of them induced a swing, of which six were “whiffs,” which works out to a 22 percent whiff rate. Of those 63 percent against San Diego, 25 of them induced a swing, but only two of those resulted in a whiff, or eight percent.
That’s a noticeable departure from the Padres game. Keller was able to induce approximately the same number of swings on the fastball, but got more Tigers’ hitters to swing through it.
But that was the only significant departure. While he threw the same number of sliders in each outing, for example, they produced nearly the same outcomes: Eight swings against Detroit versus seven swings against San Diego, both of which only produced one whiff.
Furthermore, there was no significant uptick or downtick in the velocity between each of Keller’s pitches. The most noticeable difference was his slider had a velocity of 1.1 mph faster in the game against the Tigers, but as we noted, hitters weren’t induced to swing and miss at differing rates.
Additionally, the average exit velocities only had a 0.4 mph difference, and it was actually the game in Detroit with the higher average.
Could it be the case that Keller’s fastball simply had more life on it and was therefore better at getting batters to swing and miss? Possibly. With everything we’ve seen thus far, it’s either that or the level of competition was downgraded, thus explaining Keller’s surge in production.
The final thing to note here is pitch location. There was one thing I noticed when looking at the 3D pitch chart from Baseball Savant. In the Padres’ game, Keller’s pitch map was heavily scattered, with many pitches — especially fastballs — staying/missing up, as can be seen below:
But in Detroit, Keller managed to keep his pitches to a much tighter scatter. Keller’s 11.7 percent K-BB percentage would put him down with Marco Gonzalez and Luis Castillo among starters this season, both of whom have struggled to keep runs off the board. Further, Keller’s 10.4 percent walk rate would put him 11th in baseball among starters with the highest walk rates. It could be possible that Keller’s struggles would be cleaned up with a fix on control, which we may have seen in Detroit:
Is it possible that the key to Keller’s success is control? That’s definitely a possibility — and one that’s been talked about — and it would almost certainly go a long way to providing him a path to success that many desperately want him to find.
What I am willing to conclude, however, with regard to his most recent start in Detroit is that while some of the improvements might’ve been contingent on playing an inferior team (to the Padres), much of Keller’s success during that outing can likely be attributed to his ability to control the zone, something that’s often lacking in Keller’s game.
Simple conclusion, TL;DR: It was a good/better outing.