A Simple Approach Change That Should Pay Dividends in 2023

Changeups hold a special place in the Twins’ history. Whether it be Frank Viola’s sweet lefty fade, Brad Radke’s stubborn willingness to throw his changeup in any count, or the disappearing magic of Johan Santana’s circle-change, some of the best-offspeed artists in the game’s history have called Minnesota home.

In the past decade-plus of Twins baseball, the embrace of the changeup has waxed and waned. Paul Molitor’s pitching coach, Neil Allen, was a major proponent and tried to instill a changeup emphasis throughout the organization. Then the Falvey regime had more of a fastball-breaking ball emphasis in their early years. Now, the Falvey regime seems to have pivoted back to embracing changeups (and their brethren, splitters), actively targeting pitchers who lean on those pitch types in their player acquisitions over the past two seasons (see Chris Paddack, Emilio Pagán, Tyler Mahle, and Pablo López).

Baseball Savant

From the outside, it stands to reason that this front office, with its relentless commitment to finding value in player acquisition and development, would find itself back onto changeups and splitters.

At their most basic, these pitch types are about deception. Offspeed pitches are about making batters think the pitches are something they’re not (fastballs). Where curveballs and sliders feature more easily observable and measurable movement, and fastballs feature easily measurable traits like velocity and spin, untangling the explanations for why changeups work is more difficult.

Despite all the technological advancements that have been made to measure the physical characteristics of pitches and their behaviors, it’s still very hard to objectively measure deception, and figuring out how to do so is a very active field of research. Despite the progress we’ve made, there are still some pitchers (and pitches) that outperform the physical characteristics we can measure and analyze.

Because changeups and splitters typically rely on deception from arm speed, tunneling, or funky spin that causes unusual movement profiles, or other factors entirely that we don’t yet understand, it is likely the case that the analytic models teams use to evaluate players might be underappreciating (or at least not fully accounting for) offspeed pitches. If that were true, it might be exactly the kind of thing this front office would be interested in.

López is a case in point. Many of the public-facing pitch quality models assess his changeup as average or worse based on the measurable physical traits they use as inputs. Yet, opponents have hit just .218 and slugged .360 against the 2,417 changeups López has thrown in his career. Collectively, those pitches have prevented 7.1 runs more than average per Statcast.

The models are probably missing something.

The Twins have thrown changeups and splitters at a top-ten rate each of the past three seasons, per Statcast. The results have been mixed as you can see in the table below:

Baseball Savant

Traditionally, offspeed pitches have been deployed primarily against opposite-handed batters (i.e, right-handed changeups against left-handed batters). The general philosophy is that pitches that move away from the batter are harder to hit than pitches that move toward the batter. This bears out in the data. At the league level in the Statcast era, changeups and splitters thrown by opposite-handed pitchers from the batter accumulated -33.7 runs (-0.006 per 100 pitches) relative to average, whereas same-handed offspeed offerings have cost 347.3 runs (+0.21 per 100).

In 2021, the Twins’ poor changeup performance was counterintuitive. Against same-handed batters, the Twins’ offspeed performed well enough, with -2.4 runs. It was against opposite-handed batters, where you would expect those pitch types to play well, that they struggled to a +30.4 run value.

Digging deeper, though, it’s clear that performance came from a handful of dreadful individual starting pitcher performances, including J.A. Happ (+6.8), Matt Shoemaker (+2.1), Randy Dobnak (+4.3), Charlie Barnes (+4.3), an injured Kenta Maeda (+3.1), and a miscast as a starting pitcher Griffin Jax (+10.7). It’s hard to take much team-level meaning from data like that.

Last season, though, the script flipped back to what we’d expected. The Twins’ opposite-handed offspeed pitches were worth -14 runs. What got my attention, though, was their performance against same-handed batters, especially right-on-right. At the team level, right-handed offspeed pitches against right-handed batters yielded +21.2 runs from average alone, easily the worst total in the majors and about 50% more than the 2nd-worst +14.2 total from Kansas City.

Whereas the offspeed issues in 2021 were primarily from replacement-level starting pitchers, last year’s right-on-right changeup challenges came in large part from the Twins’ late-inning relievers.

Baseball Savant

Yes, Chris Archer (+4.9) and Dylan Bundy (+1.9) contributed, like Happ and Shoemaker before, but I was surprised to see Pagán (+4.1), Jhoan Duran (+3.2), Jorge López (+1.7) factor so prominently on the right side of this chart. Even slider-heavy Griffin Jax was on the wrong side of this.

It makes sense why starting pitchers, who typically face batters more than once, might be more inclined to throw same-handed changeups to occasionally mix things up and keep their opponents guessing. But in the bullpen, the name of the game is usually about optimizing the approach and using your best two pitches as often as possible. When the Twins’ best relievers have the platoon advantage, a changeup or splitter is highly unlikely to be their best choice (perhaps save for Jovani Moran and his changeup).

To that point, here’s the same chart, but for right-on-left offspeed pitches:

Baseball Savant

Here Duran, López, Jax, and Pagán are all on the left side of the chart. Their offspeed offerings are good pitches, they just should never be deployed against same-sided hitters.

Instead, all of them should be throwing their breaking pitches against righties. That’s especially the case for Duran (-6.9 runs against RHB on his curveball) and Griffin Jax (-5.2 runs against RHB on breaking balls), but also the case for López (-0.7) and Pagán (-1.0).

It’s one thing for starters like Joe Ryan, Pablo López, and Tyler Mahle to be willing to deploy their offspeed pitches against hitters other both sides of the plate. But it’s another in late-game situations where the margins get thinner and the optimal approach is more obvious.

Hopefully, we see very few right-on-right offspeed pitches from the Twins’ best relievers in 2023.

---

John is a writer for Twinkie Town and Pitcher List with an emphasis on analysis. He is a lifelong Twins fan and former college pitcher. You can follow him on Twitter @JohnFoley_21.