Dissecting Dallas Keuchel

I’ll apologize at the outset for a lengthy post, but I do believe that the Twins’ FO and coaching staff are going to be having discussions revolving around this topic, and I feel TD should have it, too.

Causes of Keuchel’s Decline

To say that I was skeptical of Dallas Keuchel’s ability to be effective as a Twins starter would be an understatement. I definitely fall on the “Keuchel’s AAA numbers are smoke and mirrors” side of the debate, and (24hrs ago) I would have rather seen a Varland/Winder/BP day start than Keuchel taking the ball. In my opinion, most of his issues related to a lack of the precise command that won him the AL Cy Young in ’15.

In that season, his sinker averaged 90.0mph, and in 2016 (the first year statcast recorded vertical movement), it averaged 25.9″ of drop (4.1″ above average). In 2022, his average sinker velocity was 87.8mph with 30.1″ of drop (3.8″ above average). This kind of regression is normal– Clayton Kershaw’s average four-seamer has lost 3.0mph in the same 2015-2023 span.

Some of the off-speed stuff is not quite there. Keuchel’s slider has regressed heavily. It averaged 52.2″ of drop and 10.1″ of sweep at 78.5mph in ’16 while presenting a very hittable 45.3″ of drop and 9.8″ of sweep at 77.0mph in ’22. His changeup was very heavy in arm-side run (18.3″ in ’15), and he has since traded ~4″ of arm-side run for ~8″ of drop.

Here is a screenshot of his Baseball Savant page with the information about his pure “stuff” (left) and control (right):

The information on the right paints a good picture of what is occurring. In ’15 and ’17, Keuchel prevented over 40 runs with pitches on the edges. His style revolved around pitching the hitter into two-strike counts and then making them swing at a pitch they weren’t looking for on a fringe. That being said, the 1st Pitch Strike Percentage and Zone Percentage dropping 4.9% and 5.8%, respectively, since 2016 is quite worrying. He found himself in 3-0 counts 5.1% of the time in ’22 (up from 3.0% in ’15) and got to 0-2 in 19.5% of his PAs in ’22 (down from 22.3% in ’15). A changeup on the low outside corner is much more dangerous in 0-2 than it is in a 3-0 count, as the hitter is likely rolling that pitch over if he swings, yet he doesn’t have to on 3-0. Pitching into good counts is how the Keuchels/Maedas/Civales of the baseball world make their living. Keuchel has not done a lot of that since the shortened COVID season and it’s been his downfall.

His inability to work counts in his favor allowed hitters to be much choosier and sit on pitches, which is exasperated by the fact that he only throws in the high 80s. He can’t throw an F-you fastball at the numbers the way Duran, Lopez, or Jax can throw on 3-0. Getting into favorable counts is imperative and a key sign of whether or not Keuchel will be able to revive his career.

His Twins Debut 

On a surface level, Keuchel was great in his debut. Pitching 5.0 innings of 1 ER ball is about as good as anyone could ask for, especially when the leverage corps all had a day of rest. If we go any deeper, however, we start to run into some complexity. He gave up 8 hits and 2 walks (2.00 WHIP) while striking out none (4.46 FIP). Of course, the 61.9% groundball percentage is much closer to his ’15 GB rate (61.5%) than his ’22 rate (50.7%). His hard-hit percentage (28.6%) is lower than any full-season mark of his, as well as his average EV (81.7mph). To see whether the hits were lucky (on the part of the hitters), let’s take a look at every PA with Keuchel on the mound.

Top 1

Ketel Marte: Flyout (84.9mph at 47 degrees) .020 xBA      Corbin Carroll: Walk (N/A) N/A      Tommy Pham: GIDP (97.0mph at 0 degrees) .420 xBA

Top 2

Christian Walker: Groundout (39.6mph at -57 degrees) .220 xBA      Lourdes Gurriel: Single (88.5mph at 16 degrees) .930 xBA      Emmanuel Rivera: FC (83.1mph at -26 degrees) .070 xBA      Jake McCarthy: Single (59.4mph at -55 degrees) .330 xBA      Nick Ahmed: Groundout (96.9 mph at -3 degrees) .330 xBA

Top 3

Carson Kelly: Single (64.5mph at -1 degrees) .120 xBA      Ketel Marte: GIDP (78.4mph at 6 degrees) .290 xBA      Corbin Carroll: Single (98.6mph at 7 degrees) .590 xBA      Tommy Pham: Walk (N/A) N/A      Christian Walker: Lineout (112.1mph at 8 degrees) .780 xBA

Top 4

Lourdes Gurriel: Single (88.4mph at 18 degrees) .760 xBA      Emmanuel Rivera: Flyout (94.0mph at 40 degrees) .020 xBA      Jake McCarthy: Flyout (75.2mph at 54 degrees) .100 xBA      Nick Ahmed: Lineout (97.0 mph at 17 degrees) .490 xBA

Top 5

Carson Kelly: Double (103.5mph at 5 degrees) .590 xBA      Ketel Marte: Groundout (80.4mph at -8 degrees) .120 xBA      Corbin Carroll: Single (60.2mph at 30 degrees) .520 xBA      Tommy Pham: RBI Groundout (94.6mph at -16 degrees) .120 xBA      Christian Walker: Single (26.8mph at -42 degrees) .250 xBA     Lourdes Gurriel: Groundout (93.0mph at -16 degrees) .140 xBA

We can use an xBA of .350 to be our threshold of “allowable” xBA, where a hit recorded with an xBA less than .350 is judged to be “bad luck” and an out recorded on a ball with an xBA above .350 is “good luck.” Using this threshold, Keuchel had three instances of “good luck” and three instances of “bad luck.”

His average xBA was 0.343 with two walks, though only one of the eight hits went for extra bases (and none of the “loud outs” would have gone for extra bases except maybe the Walker lineout). An average xBA above .300 is quite atrocious, as it is considerably higher than his .289 xBA mark in 2022, when he was among the bottom 1% of the league in this category. Sample size and the lack of strikeouts are likely the culprit of this large number, though a 15% K-rate would only reduce the xBA to a .292 mark.

In terms of working the count, he reached zero 3-0 counts and had the hitter on an 0-2 count 21.7% of the time (very close to his ’15 number). In this regard, he performed quite well and was able to get big outs by getting ahead of guys early. Additionally, the McCarthy hit in the 2nd, the Carroll hit in the 5th, and the Walker hit in the 5th all had xBAs above .250 despite being hit slower than 61mph. Expected batting average incorporates the batter’s footspeed, and three balls hit under 61mphs resulting in hits is bad luck (despite what xBA says) as the pitcher has won if he can induce extremely weak contact.

Conclusion

Obviously, it’s hard to make concrete conclusions about a five-inning start. We’re still debating how effective Pablo Lopez, Griffin Jax, Sonny Gray, and Emilio Pagan have been, and they have had 69% (nice) of a season to give us data. As it pertains to Keuchel, the early returns are promising. He’s using the same pitch mix (sinker/changeup/cutter/slider) that he’s used for his whole career and has worked the count in a way we haven’t seen from him since his ’15-’17 stretch of dominance. His xBA allowed is quite worrying, though three soft singles likely skewed this small sample of data. His groundball rate has returned to the 60%+ range, and he only gave up two extra-base-worthy batted balls in 5.0 IP. On the whole, his performance doesn’t raise a ton of additional concern from me; however, it hasn’t inspired a ton of optimism, either.

I would feel really uncomfortable if Keuchel had to start a playoff game, but as a 6th starter, I think he brings some promise. What do you guys think? Does he deserve to start another game? Was his debut a fluke, or was it a sign that he’s a changed pitcher? Thanks for staying with me through this statcast-infused salad, and I hope we can have a good discussion about Keuchel’s future with the franchise.