Why did Kenta Maeda throw more fastballs?
A few weeks ago, my Pitcher List colleague Casey Drottar looked at the struggles of Twins right-hander Kenta Maeda last season. After Maeda challenged for the AL Cy Young in his freshman season at Minnesota, he endured a Murphy’s Law-type season last year, battling a nagging groin injury and inconsistent performances on the way. of a 4.66 ERA in 106.1 innings before the big hit — a major elbow injury that required reconstructive surgery — ended his season early. Now Maeda is expected to miss the majority of the 2022 season.
Given that Maeda is unlikely to be a big factor this season, it might seem odd that we’re analyzing him twice in a month. However, like a lot of good baseball writing, Casey’s article, which focused on Maeda’s fastball, got me thinking.
It’s been well documented, including by Cole Bailey here at Pitcher List, that much of Maeda’s stellar 2020 campaign was made possible by a significant adjustment to pitch blending that saw him introduce his slider ( 38.7% usage) and his division change (29.3%) more often than his fastballs (28.5% four-seam, sinker, cutter combined) and more often than he had previously with the Dodgers.
Height-level performance data supported this adjustment. According to Statcast’s running values, Maeda’s slider has long been his best throw, generating above average results every season since 2017. Pitchers throwing their best throws more often have been all the rage in recent seasons, Maeda has jumped on board that trend and had his best season as a starting pitcher yet.
But, as Casey noted, he went back to relying more on his fastball last season, especially when he fell behind in the count:
Created by @visual_endgame on Twitter.
My question is why? The Twins have been one of the most ball-dependent organizations the past two seasons, including when their pitchers are behind the count. Maeda’s pitch-level performance data from his excellent 2020 show that his fastballs performed better when he threw them less often. Given their general ethos, it seems unlikely the Twins will take this new data point to mean they should go back to their previous ways and rely more on fastballs again. This would go against their organizational strategy and the plan they had for Maeda when they targeted him in the trade. To me, it seems obvious that something else had to happen. But what?
I decided to start by looking at his overall ability to throw punches and move up the count. In general, pitchers command their fastballs better and we saw above that Maeda went to his fastball more often when he was late last season. Perhaps he was experiencing less control of his breaking and speed throws, driving him into unfavorable counts where he really needed to throw quality strikes?
Disconcertingly, Maeda threw more shots into the strike zone last season (45.2% zone rate) than he did in 2020 (43.8%). At the same time, however, Maeda has fallen behind in the count to start at bat more frequently than he ever did before last season. His first pitch strike rate in 2021 was just 59.8%, down from 64.5% the previous season, and just below his career average of 62.9%.
Partly because of this, Maeda threw the lowest pitch percentage while leading in the count (28.3%) and the highest pitch percentage while trailing in the count. tally (27.4%) of his career. In both cases, these rates were about three percentage points lower than the previous season.
It is known that Maeda turned to his fastballs more often when he was behind. I was curious if he was also choosing to work more with his fastball to start scoring, which could indicate a change in strategy rather than a change in circumstance.
The table below shows the pitch types selected by Maeda in 0-0 over the past two seasons:
In general, Maeda performed quite similarly to starting board appearances last season. In his two years as a twin, he threw a first pitch other than a fastball 55-56% of the time, about half of those were sliders.
That suggests that, at least in terms of strategy for starting a hitter, he took the same approach last year.
We already know he didn’t throw as many strikes on the first pitch last season, so he must not have executed this approach as well as he did in 2020. If we generally assume that a pitcher intends to throw a strike on the first pitch at bat, looking at 0-0 called rates can be informative in understanding if a pitcher is having trouble executing a preferred strategy.
These data are quite revealing. Maeda’s called rates on first pitches were noticeably up on all of his pitch types except his split change. This was especially true for the sliders he used most often to start the count, which were up eight percentage points from 2020.
If Maeda was having trouble executing his slider the way he wanted, it could have led to him returning to his fastballs once he fell behind in the count. After all, you wouldn’t blame a pitcher for going to another pitch after often missing like this on 0-0:
To be fair to Maeda, I select illustrative examples here. All pitchers will have occasional misfires which will result in pitch pitches like those in the video above. However, you can tell from Maeda’s body language after some of these pitches that he’s frustrated with his cursor control.
His global field location data backs up the idea that he was right. Compared to 2020, Maeda’s slider locations were much more variable last season.
You can see that he sprayed his cursor across a wider range of locations than he had before. More of them ended up way off the plate on his glove side or left unfinished and on his arm side than in 2020 when he did a better job spotting them closer to the edge down and to the side of his glove.
I should note that order evaluation is always a tricky proposition due to our lack of location intent data and the sample sizes with the two seasons are different. Maeda launched 270 more sliders last season than in the shortened 2020 season, which could skew the visuals. To test if his mastery of the pitch was really more variable last season, I ran the summary statistics of the detailed pitch-by-pitch location data available from Statcast and calculated the standard deviations of the plate location data. Maeda’s sliders as a crude measure to gauge if her command was as clean as it had been:
Standard deviation of cursor locations
Last season, the standard deviations of the X and Z plate locations of his sliders were noticeably higher than they had been in 2020 and F-tests show the differences between them to be statistically significant. This confirms, as the cards indicate, that he located his cursor less consistently last season.
Pitch movement and release points
A potential explanation for Maeda’s difficulty spotting her cursor is a noticeable change in the terrain motion profile. On average, Maeda got about two inches more vertical movement (35.5 inches with gravity versus 33.4 inches) and two inches more horizontal movement (5.2 inches versus 3.2 inches) on her slider l last year than the previous season.
Perhaps more impactful than simply having more movement was that the added horizontal movement was quite variable, ranging from just an inch or two to eight or nine inches, from output to output:
Some variation from game to game is natural and expected, but that variability seems high. Apparently every time he threw his cursor it had a different pause. Given that, it’s no wonder he had trouble consistently locating it where he wanted it.
Here it should be mentioned that Maeda suffered a lingering groin injury in early May for a few rounds before spending three weeks on the injured list until June. It’s hard to know for sure how much weight to put on the injury affecting his ability to throw his shots the way he wanted, but he admitted it caused him to alter his mechanics. The data suggests that it had negative effects.
The extra horizontal movement may also explain why he moved his starting spot on the pitcher’s rubber to third base mid-season (which Casey noted). It’s a logical adjustment to make if he had more horizontal break and missed the plate on his glove side more often than he wanted to. (It’s also the one the Twins had done in a similar situation with José Berríos a season earlier).
Ultimately, this all makes up a plausible explanation for why Maeda strayed from his slider-heavy scheme and back to his fastball, especially in the batter count. Given this data, this change seems to be the result of an inability to execute rather than a deliberate strategic choice to present his fastball. Either way, when he returns to the mound at Target Field, how he controls his slider will be a key thing to watch.
Photos by Icon Sportswire | Adapted by Doug Carlin (@Bdougals on Twitter) | Visualization of Fastball frequency data created by @visual_endgame on Twitter.