Trouble with the curve (and slider... and cutter...)

Among the most popular story lines illustrating the Twins’ disappointing start to this season is the suddenly vulnerable pitching staff, particularly the bullpen. Through 23 games, Minnesota hurlers have combined for 4.24 ERA (19th), 4.27 FIP (20th), and just 1.2 fWAR (t-22nd).

Zooming in specifically on the relievers, those numbers dip to 4.83 ERA (27th), 4.19 FIP (18th), and 0.0 fWAR. In addition to those replacement level run prevention numbers the relievers have already taken 8 losses (most in MLB) and converted just 4 of 9 save opportunities.

These stats, while still very much coming from a small sample, paint a starkly different picture than the performance of the Twins pitching the past two years. Pitching arguably carried the 2020 team and the past two seasons the staff ranked 3rd and 2nd in MLB in fWAR, respectively.

Combining the two seasons, the Twins ranked 1st in WAR (33.8), 4th in FIP (3.96), and 7th in ERA (4.02). The ‘pen was a strength of the 2020 Twins and the team’s relievers tied for the major league lead in FanGraphs’ version of WAR (3.6). Twins’ relievers finished sixth in ERA (3.62) and fifth in FIP (3.85) and strikeouts per nine innings (10.36).

So, what’s going on?

There are a few signs that the pitchers have experienced poor luck that should even out some as the season continues, especially with the relievers. The disparity between the bullpen ERA and FIP above (0.64, 4th-largest) is worth nothing. So is the relievers’ collective 61.6% left on base percentage, which is tied for the lowest mark in baseball. Also supporting the case for optimism is that the Twins’ relievers have struck out 27.0% of the batters they have faced and walked just 8.8%. The delta between the Twins bullpen’s K% and BB% is 18.2%, a gap that ranks 6th-best in MLB.

Beyond luck though, there is a factor that is within the players’ control that has emerged as a key source of the pitching issues. I wrote this winter that a major piece of the Twins’ success on the mound the past two seasons had been a strategic change that included some of the heaviest deployment of breaking balls in baseball. That reliance on spin has continued this season. The combination of sliders and curveballs makes up 35.8% of Twins pitches thrown this season, the 7th-highest rate of breaking balls.

That’s a slight decrease from last year’s 40.6% rate, but it’s more likely that shift is due to personnel changes than any kind of organizational philosophy change. Swapping out curveball specialist Rich Hill and slider-ballers Sergio Romo and Matt Wisler from last year’s roster for the fastball heavy J.A. Happ, Hansel Robles’ split-change and slider, and Alexander Colomé’s cutter could explain that change in the pitch type data. If you consider Colomé’s cutter a breaking ball (he deploys it as one) then the club’s breaking ball rate is 45.1% so far in 2021. That would rank as the second highest rate of breaking pitches in MLB.

While the rate of breaking balls has stayed similar, the effectiveness and results against those pitches have not. Collectively, the league hit .196 / .256 / .329 against Twins’ breaking pitches in 2020, good for a very weak .253 wOBA. As measured by FanGraphs’ version of pitch type linear weights, the club pulled off the rare double feat of having baseball’s most valuable sliders and most valuable curveballs in 2020. By Statcast’s version of pitch type run value, Twins’ breaking balls were the second best last season.

So far in 2021, Twins’ breaking pitches have been among the least effective. The combined triple-slash line against them is .229 / .276 / .435, good for .314 wOBA. In comparison to last season, the batting average and on base percentage numbers are not terrible, but the slugging percentage indicates these pitches have been getting hit hard. That’s supported by the fact that Minnesota pitchers have already allowed 14 home runs on pitches classified as breaking pitches by Statcast, which is tied with Atlanta for the most in MLB. FanGraphs’ pitch type linear weights numbers track similarly to the Statcast numbers — Twins’ sliders rank 28th, curveballs 22nd, and cutters 29th, by those measurements.

This early in the season, the team level stats can still be influenced significantly by individual performances, and that is largely what is happening here. Let’s take a look at the pitchers having the most trouble with their breaking pitches:

Alexander Colomé’s Cutter (+7.2 runs, per Statcast)

We’ve talked about Colomé’s struggles a lot on the virtual pages of Twinkie Town this season. I won’t rehash the gory details, other than to say again, Twins’ cutters rank 29th in baseball by Statcast run value and have allowed a .300 batting average, .567 slugging percentage, and 2 home runs. It’s worth noting that Alexander Colomé is the only Twin that throws a cutter, so he’s shouldering this one all on his own. In fairness, cutters aren’t as ubiquitously used by pitchers as fastballs and sliders, but the fact that Colomé’s cutter by itself is worse than 28 other team’s combined cutters would be hilarious, if it weren’t so disappointing and sad.

I wish there were some advanced stats I could point to that suggested his cutter should come around, but the expected stats, exit velocity, raw spin, and movement data suggest the pitch has largely deserved it’s performance thus far. It’s not moving as much as it did last season (both vertical and horizontal movement is down noticeably) and Colomé’s command of it has been poorer than before. Until he can re-discover his feel for the pitch, it’s going to be very hittable, like the one he served up to Jordan Luplow to end the game earlier this week in Cleveland.

Caleb Thielbar’s Curveball (+5.3 runs)

Thielbar has thrown all of 29 curveballs this season, six of them have been put in play, and two of those have ended up in the outfield bleachers. Last season, just six of 94 Thielbar curves were even put in play, he allowed no hits off the pitch, and it was worth -3.0 runs by Statcast. Batters have hit the lefty’s curve to the tune of .455 / .455 / 1.000 so far this season and the expected stats are not much better — .369 / .369 / .806.

But it’s only been 29 pitches, which is a tiny sample. On the plus side is Thielbar’s 46.7% whiff rate on the pitch. It seems when it’s not getting hit hard, it’s getting swings and misses. To me, this appears to be a case of getting hurt by a few bad locations early in the year, shown below:

baseballsavant.mlb.com

The highest in zone pitch data point in the plot above resulted in Justin Upton’s grand slam a few weeks back in Anaheim:

It is worth noting Thielbar is throwing his curveball with slightly less extension and with about 100 RPM of raw spin less than he did last season. Those could both be impacts from the cold, early season weather. Those bear watching, along with the location of his curveballs, as the season progresses.

Kenta Maeda’s Slider (+4.9 runs)

Maeda rode his slider hard last season, using it for 38.3% of his pitches. So far in 2021, he’s doubled down on that approach, offering his slider 43.2% of his offerings. Maeda has clearly been fighting his mechanics and command through his four April starts, which is evident in his body language on the field and his 60.7% first pitch strike rate (down from 64.5% last season). It is also evident in his overall pitch location data, especially with his slider. Below are two plots comparing Maeda’s slider locations last season to his slider locations so far in 2021:

baseballsavant.mlb.com

baseballsavant.mlb.com

You can clearly see in the second plot, that Maeda has missed up in the zone with his slider more this season. Another telling sign is the number of sliders that are out of the zone, up, and to his arm side (right handed batter side, reader’s left hand side of the second plot). Maeda’s sliders should break down and away from right-handed batters. The number of times Maeda has thrown non-competitive sliders up and to his arm side is a clear indication his mechanics are off.

The result is a 6.14 ERA that has been inflated by a whopping 7 home runs allowed in just 112 batters faced. Four of those homers have come from sliders and batters have hit .386 and slugged .773 against Maeda’s primary pitch. The four homers, as you might suspect, have come on sliders that missed up, arm side, and in the heart of the plate:

baseballsavant.mlb.com

Interestingly, the raw spin characteristics on Maeda’s slider are largely the same as last season and he is actually getting more movement on the pitch (about an inch and a half horizontally and almost 3 more inches vertically) so far in 2021. Those are indications to me that Maeda’s struggles with the pitch will cease if and when he finds his consistent mechanics that limit these kind of misfires and location mistakes.

---

John is a contributor to Twinkie Town with an emphasis on analytics. He is a lifelong Twins fan and former college pitcher. You can follow him on Twitter @JohnFoley_21.