The 2019 Twins had a very effective bullpen, leading the major leagues in such advanced metrics as relief Fielding Independent Pitching (3.92), finishing third in total fWAR by relief pitchers (7.4) and third in relief K%-BB% (17.6%). To some degree, an argument can be made that the Twins relievers, despite their overall success, were a little unlucky in 2019. Their combined ERA (4.17) exceeded their fielding independent estimate by 0.25 runs – the 7th largest positive gap in the MLB – a gap that was fueled largely by the second highest batting average allowed on balls in play (BABIP), 0.317.
All told, the Twins relievers made 524 appearances in the 2019 season, covering 573.2 innings. Something stands out when you dig a bit deeper. Of these appearances and innings pitched, the large majority of them were taken by right-handed pitchers. Take a look.
Right-handed pitchers accounted for more than 82% of the Twins’ relief appearances and about 77% of relief innings pitched. The left-handed appearances are dominated, unsurprisingly, by Twins closer Taylor Rogers. He accounted for two-thirds of the team’s lefty appearances (60), and more than half of the team’s lefty relief innings pitched (69). Another 45.2 innings pitched by lefty relievers came in just 18 games from a combination of left-handed starters Martin Perez, Lewis Thorpe, and Devin Smeltzer working in long relief.
Despite that right-hand dominance, the Twins bullpen performed almost equally as well against batters from both sides of the plate.
In terms of batting average allowed and on base percentage allowed the Twins relievers were about as equally effective as you can be. What’s interesting is the numbers above relative to left-handed batters are among the tops in baseball. The 3.65 FIP against left-handed batters number ranked 2nd. The 3.75 ERA, fifth.
This goes against conventional wisdom. Traditionally, it is well understood and researched that batters perform worse against same handed pitchers (i.e. right-handed batters perform worse against right-handed pitchers and vice versa). This is known as “the platoon split” and has been extensively studied and deployed in games throughout baseball history. One of the foundational baseball analytics treatises - The Book, Playing the Percentages in Baseball by Tom Tango, Mitchel Lichtman, and Andrew Dolphin, published in 2006 - estimated that the average platoon advantage is worth 44 points of wOBA. In simpler terms, that means an average batter or pitcher will be 44 points better when facing an opposite handed opponent.
The Twins bullpen performance against lefty batters last season is then even more impressive, given they were most often working without the platoon advantage and did not supplement those right handers with the soon to be extinct left handed-one-out-guy (LOOGY). For last season at least, it seems the franchise was able to put together a group of right-handed pitchers that could get both right-handed and left-handed batters out. While the exact makeup of the 2020 season schedule and rosters is still unclear, we do know which pitchers the Twins currently employ and had invited to 2020 big league camp. Those that might spend significant time in the big league pen are again primarily right handed, although the team did seem to try to stockpile some veteran left handed depth on minor deals this winter (Danny Coloumbe, Blaine Hardy, old friend Caleb Thielbar). This depth took a hit earlier this week when we learned Blaine Hardy has undergone Tommy John surgery. The roster data below is available at Fangraphs using the depth charts from RosterResource.
Let’s breakdown the big league pen, taking a look at how well equipped they are to handle lefties and navigate the new three batter minimum rule. While Rogers returns after playing a huge role last season, Tyler Duffey and Trevor May both have had up and down swings season to season before, Zack Littell and Cody Stashak were up and down frequently between Minnesota and Rochester last year and may not yet be considered “proven”, Sergio Romo was a mid-season addition from Miami, and Tyler Clippard (Cleveland) and Matt Wisler (San Diego and Seattle) played for different teams. For this analysis, it’s worth noting that I’m only working with pitchers that figure to be exclusively relievers. That leaves out guys such as Lewis Thorpe, Rich Hill, Devin Smeltzer (all lefties), Jhoulys Chacin and anyone else who may find their way into the bullpen at points this coming season. The Twins likely prefer them to be starters, so it’s cleaner to exclude them here.
How might this bullpen group fare? First, let’s look at this group’s splits from 2019 compared to the American League overall. For this, I’m using weighted on base average (wOBA) as the statistic, which is one of the most popular catch-all offensive statistics. It’s shown on the same scale as on base percentage and based on the simple concept that not all hits (and other ways of reaching base) are of equal value.
A quick glance, unsurprisingly, shows this group of Twins performed well compared to the average American League reliever in 2019. A couple of key things jumped out for me.
First, it seems that Rocco Baldelli has many potential options to be successful against right-handed batters (top blue chart) – Rogers, Romo, May, Duffey, and Stashak are all significantly better than league average against right handed batters.
Second, the left-handed Rogers fared better against righties than lefties last year – though his numbers vs. lefties were more than respectable and have been his stronger side throughout his career. His ability to get hitters from both sides of the plate out is likely a major factor in the Twins being comfortable having a left-handed closer, something that has not been particularly common throughout baseball history.
Third, including Rogers, the Twins have a strangely high number of relievers with reverse platoon splits. Only Romo, Stashak, and Wisler fit the traditional mold of being better against same handed opponents and Littell’s splits differences are small enough to be negligible. The data shows Clippard, Duffey, May, and Rogers, while clearly effective overall, are stronger against opposite handed batters.
Fourth, it seems Rocco also has several good options against lefties, led by those right handed reverse splits guys, Clippard, May, and Duffey. Clippard in particular, looks to be exceptionally difficult on left-handed batters. That 0.186 wOBA vs. lefties was third best among all relievers that pitched 20 or more innings last year.
Since the focus of this post is about whether another left-handed reliever is needed, let’s dig into some Statcast data from baseballsavant.mlb.com on Clippard, May, and Duffey to see how each of these guys gets that reverse advantage against lefties.
#36 – RHP - Tyler Clippard
Clippard is primarily a changeup-fastball pitcher, and he throws two variations of each. The charts below show his pitch mix and movement profiles. His fastball velocity overall is well below average (just 13th percentile), but he is able to get excellent velocity differential of 8 to 12 mph with the changeup and split-finger that he throws more than 55% of the time combined. The pitch movement chart is from the catcher’s point of view and the vertical line bisecting the chart implies a straight pitch with no horizontal movement. The top of the chart implies a pitch with no vertical movement. All pitches have some vertical movement downward due to gravity.
The data shows that Clippard features a couple of interesting things. First, he’s one of the rare birds that throws both a changeup and a split. Most of his pitches feature movement downward with arm side lateral movement (away from lefty batters). What stands out most to me is the amount of vertical movement on the changeup and the split, but for opposite reasons. The split (teal color), with 42 inches of vertical break on average led major league baseball in 2019. The changeup (green) on the other hand, had some of the least vertical break in the major leagues - 29.1 inches, about 6 inches less than changeups thrown with similar velocity.
See how close together the orange and green circles are on the right side chart above? And how far apart the red and teal circles are? The way these pitches pair together - changeup with sinker; split with four seam fastball - gives Clippard multiple options to attack hitters. The changeup’s lack of vertical drop makes it look like his sinker, but 11-12 mph slower. The split’s violent movement downward pairs very well with his four-seam fastball up in the zone because of the wide difference in movement and about 8 mph of velocity change.
For Clippard to maximize his arsenal, we would expect to see him work in and around the same parts of the strike zone with his sinker-changeup combo, and to work in different and opposite parts of the zone for his four-seamer-split combo. And that’s exactly what he does:
These plots are again from the catcher’s point of view, and we can clearly see Clippard working his arm side of the plate with changeups and sinkers, pretty evenly distributed up and down. On the other side we see a clear pattern of fastballs up and splitters down. These combinations make it exceptionally difficult for left handers in particular to pick up as they must account for the full strike zone top to bottom and contend with large velocity changes from pitches that look and move very similarly. It’s no wonder his career wOBA allowed vs. lefties is only 0.260.
#21 – RHP - Tyler Duffey
Duffey has been a different guy since he was recalled to the majors last April. Since, he has basically become a two-pitch pitcher – throwing his four-seam fastball and slider about 89% of the time and roughly equally against left and right-handed batters. This change was by design – intended to maximize the amount of time his fastball and slider (some sites call his slider a curveball – I’ll use slider for this post) looked the same out of his hand on their way to home plate, a concept called “tunneling”. This idea may be best explained with a GIF:
Kerry Wood, Slider/Sword & running Fastball, Individual Pitches + Overlay. pic.twitter.com/HoJ3cfioPu— Rob Friedman (@PitchingNinja) March 28, 2020
That’s Cubs right hander Kerry Wood during his 20-strikeout masterpiece of the Astros in 1998 and you can clearly see his fastball and breaking ball leave his hand on the same plane and trajectory, only to break in vastly different directions as they approach home plate. The longer the two pitches stay together after release, the harder they are to pick up for the batter. Duffey’s evolution has allowed him to do something of the same with his two best pitches.
Duffey had great success against hitters from both sides in 2019. His four-seam fastball only has a small amount of arm side run (5.9 inches on average, less than MLB average at his velocity) and it has solid, but not spectacular velocity (67th percentile), just under 94 mph on average. Even so, the fastball was effective in 2019, yielding only a 0.252 wOBA.
But it’s the slider that stands out most for Duffey – particularly for it’s vertical break. In 2019, Duffey’s slider averaged 50.7 inches of vertical drop (gravity included), nearly 13 inches more than the average sliders with similar velocity. That additional 13 inches was the most in the major leagues in 2019. As a result, the slider was immensely successful against hitters from both sides, yielding only a 0.169 batting average allowed and a 0.227 wOBA overall. Naturally, with that much vertical movement you would expect to see Duffey working up in the zone with his fastball and down below the zone with his slider in order to maximize that tunneling effect – which is exactly what he does:
In Duffey’s case, the pairing of these two pitches is highly effective against hitters of both sides of the plate – essentially negating the need for another pitch that breaks away from a left-handed hitter. When the breaking ball is located below the zone, off the fastball up, it’s just plain hard to hit right-handed or left-handed. Just ask the Tigers’ Jordy Mercer:
Tyler Duffey, unfair 86mph Curveball. pic.twitter.com/t7JH1zCPC0— Rob Friedman (@PitchingNinja) August 31, 2019
#65 – RHP - Trevor May
In Trevor May’s case, it begins with the four-seam fastball. May has more velocity than his bullpen buddies averaging 95.5 mph (85th percentile) and he uses that big heater often – more than 61% of his pitches. This usage is consistent against both right-handed batters and left-handed batters. For off-speed pitches, the former starting prospect has 3 offerings that he deploys between 10% and 15%, depending on the handedness of the batter. The movement profiles on May’s pitches show us that he is equipped to utilize the same approaches as Clippard and Duffey.
A couple of things to note on the movement chart. First, May’s changeup (green) has a comparatively small amount of vertical drop (even less than Clippard). See how close the green circle is to the red circle? His changeup averaged 21.7 inches of vertical movement, which was 6.8 inches less than changeups thrown around 87-88 mph in 2019 (4th least in MLB), but still managed about 8 mph of velocity separation from his fastball. So the changeup moves similarly to his fastball, but slower. And the slider (yellow) has a comparatively large amount of vertical drop – 43.2 inches on average, about 5.7 inches more than comparable sliders thrown around 84 mph. Like Duffey, that additional 5.7 inches ranked in the top 15 in the MLB in 2019 and came with nearly 12 mph of velocity differential from the fastball.
And when we look at how May deploys these weapons, we find that he does so in a manner than looks an awful lot like his bullpen mates.
Like Duffey, against right-handed batters May primarily goes fastballs elevated and sliders down and away – maximizing that tunneling effect. Like Clippard, against lefties May works his fastball and changeup in and around the zone in similar locations, relying on the velocity difference and similar movement profiles for deception. And, when he wants to work beneath the zone to get movement differential from his power fastball, he can do so at an even lower speed by mixing in a 79-mph curveball.
With these three right handers and their arsenals, it seems the Twins should be just fine without another lefty to pair with Rogers in the bullpen. Rocco has three very solid options that, while right-handed, have the pitch combinations and results to give confidence they can handle left-handed batters. The Twins should be well positioned to navigate the new three batter minimum rule.
Have questions or ideas for future breakdowns? Please leave them in the comments.