We’ve now completed four seasons since the Minnesota Twins franchise hired Derek Falvey as Chief Baseball Officer and Thad Levine as General Manager after the 2016 season. At the time of their hires, the Twins had just finished an MLB worst 59-103 season that marked the franchise’s fifth losing campaign in the last six (not that any of us need to be reminded). That span followed a decade-long run of success under former GM Terry Ryan and manager Ron Gardenhire. The Twins success in the first decade of this century was built on solid pitching and much of the team’s decline in the early 2010s can be chalked up to a cratering on the mound:
For as bad as the numbers on the right side of the chart look, they are even worse when you put them into context of the overall run environment of the game in those time frames. ERA- and FIP- are more informative for comparing performance across different time periods because they are set relative to the league average. Both of these stats set league average at 100 and every point above or below 100 is a percentage point away from average. From 2001 to 2010 the Twins collectively posted an ERA- of 94 that tied for 5th best in MLB. By FIP- they were 12th best at 98. From 2011 to 2016 the club’s 113 ERA- and 106 FIP- were both far and away MLB’s worst.
When Falvey and Levine took charge, it was crystal clear what their number one priority would have to be: fix the pitching.
Now that we have four seasons in the books with them calling the shots, the results are in:
They’ve fixed the pitching. The club’s pitching has improved steadily and arguably carried the 2020 team. The past two seasons Minnesota’s pitching staff has ranked 3rd and 2nd in MLB according to Fangraphs’ version of WAR. Combining the two seasons, the Twins rank 1st in WAR, 4th in FIP, and 7th in ERA. It remains to be seen if this will continue, but what was once the Achilles’ heel of the organization has become a strength.
So, how did they do it? In this series I’m going to look at how the new baseball leadership fought this problem and dissect what the data tells us about the organizational pitching philosophy they’ve put in place.
Macro changes in pitching strategy
By the time Falvey and Levine took over the Twins, baseball was in the midst of big shifts in pitching approaches. Gone were the days of pitching to early contact and batters choking up to put the ball in play. In their place were an emphasis on the three true outcomes of home runs, strikeouts, and walks. As you’ve no doubt read before, the most notable shift across baseball the past two decades has been a steady year over year increase in strikeouts. As a percentage of plate appearances MLB’s overall strikeout rate has increased every year since 2005:
Digging in to how those strikeouts have been achieved reveals some surprising patterns. Since 2002 the league’s average fastball velocity has steadily increased, almost without pause. In that span the average major league fastball speed has increased from 89 miles per hour to just above 93 miles per hour:
Counter intuitively, though, the overall use of fastballs has decreased at a similar rate since 2002:
MLB-wide, fastballs have gotten faster and they’ve also been used less frequently. In place of those fastballs teams have increased breaking ball (i.e., curveballs and sliders) usage. The chart below illustrates the 2003 season was the low-point this century in terms of breaking balls thrown — 21.8% of all pitches according to Fangraphs’ data.
That fraction increased and then sat about 24% each season 2005 to 2015 before taking off even more over the past five seasons. The use of sliders has primarily been behind this increase, going from 11.8% in 2003 to 18.8% in 2020. Throughout, curveballs have remained constant around 9%-10%.
The logic behind this widespread change in approach is supported by clear evidence — non-fastballs yield better outcomes for pitchers than fastballs. According to an August article by Sports Illustrated’s Tom Verducci the major league batting average against fastballs is between .230 and .250, depending on the type thrown (i.e., four-seam, sinker, etc.). Against non-fastballs that drops to .211. In 2019, hitters slugged .481 against fastballs, and just .387 against non-fastballs.
Minnesota fell behind
The Twins, unsurprisingly given the results I showed you at the start of this article, were slow to catch on to these trends under the prior decision making regime. Even during their successful run through the 2000s the pitching staff was not often driven by strikeouts — they only finished with an above MLB average strikeout rate 3 times (2004, 2006, 2007). But we know those three above average seasons were due more to one Johan Santana than any kind of preferred organizational strategy. From 2011 through 2016, the club ranked dead last in the MLB in strikeouts per nine innings. The numbers make clear the organization had earned and deserved its reputation for pitching to contact.
Similarly, after being right at league average in fastball velocity in 2010 (91.2 mph) the Twins average heater velocity held steady, ranging between 90.9 mph and 91.9 mph from 2011 to 2016. That velocity ranked no higher than 21st and as low as 27th among the 30 MLB teams.
Despite the low velocity the team continued to throw them a lot relative to the rest of MLB:
The results were predictably terrible. Fangraphs’ calculates pitch type linear weights, which sum up the changes in run expectancy that occur from pitch to pitch, and can tell us how well a pitcher (or pitching staff, in this case) performs when using a certain pitch type. As the right side of the chart above shows, the Twins fastballs were among the game’s worst before Falvey and Levine took the reigns. In 2016, the season before Falvey and Levine took over, MLB hitters produced a .307 / .377 / .523 line against Twins’ fastballs, good for .384 wOBA. By comparison, then Mariners’ designated hitter Nelson Cruz posted .383 wOBA on his way to hitting 43 home runs.
The Falvey-Levine Regime
The case laid out above makes it seem obvious that a more modern approach was needed. Falvey made clear he intended to run his administration in a more modern way when he came aboard. The first sign came in his introductory press conference when he declared his administration would “root our decision making in evidence-based practices, both subjective and objective in nature.”
The second was his belief in customized player development, telling Parker Hageman at TwinsDaily, “Everything has to be individualized. I think what we need to do is find out what systems work for some of our players, what they are already doing. We need to learn what they are doing now and ask how do we build toward a vision and if that vision means a different type of arm care, or maybe a different type of velocity and growth, we’ll employ those tools for those players. I never like to shut the door on any of those systems. I like to evaluate them and see where they fit within the mix of what we are doing.”
In short, the Twins would try to optimize for the strengths of their individual players and the approach for doing that would be data-driven. How has that shown up?
Spinning Breaking Balls
Given the awfulness of Twins’ fastballs prior to Falvey and Levine taking over and the changing macroenvironment, the evidence clearly pointed to Minnesota pitchers needing to spin more breaking balls in lieu of fastballs. While that modification took a few seasons to show up at the team level, it has clearly taken hold the past two seasons:
In 2020, fewer than half the pitches thrown by Minnesota pitchers were fastballs. In their place, more breaking balls, especially sliders. The team’s curveball rate has more or less stayed steady around 13% the past four seasons. Sliders, however, have increased each season, more than doubling from Falvine’s first season in 2017. The takeaway is clear — the Twins have finally joined the spin club created by their opponents earlier in the decade.
Thinking back on Falvey’s comments about optimizing for player’s individual strengths, you can start to see a pattern with how the Twins have tried to coach adjustments to develop pitchers with a propensity to spin breaking balls. Perhaps the best example of this is now reliever Tyler Duffey. Back in 2016, Duffey was quickly becoming a failed starting prospect. In 26 starts, Duffey posted a 6.43 ERA and 1.50 WHIP while throwing 54.5% fastballs and 38.7% breaking balls. In 2017, Duffey converted to the bullpen and the results were a mixed bag (4.94 ERA, 3.72 FIP) while he leaned even more heavily on fastballs (59.5%). By 2018, he spent much of the year on the Rochester shuttle and posted a 7.20 major league ERA, throwing even more fastballs (61.7%) in the process. Despite his struggles, there were signs his breaking ball was a hidden weapon — like it’s strong 35% whiff rate and low expected batting averages and slugging percentages.
After beginning 2019 in Rochester again, Duffey returned to Target Field mid-season with a new approach. With the help of new pitching coach Wes Johnson, the new Duffey threw 45% curveballs. In 2020, he bet on that approach further, throwing the curveball 53% and fastballs just 47%. The results speak for themselves. In 81 innings the past two seasons, Duffey has pitched to a 2.31 ERA, 2.91 FIP and batters have hit just .188 against him. The breaking ball usage has enabled him to get swings and misses that he previously couldn’t. In his first year as a reliever Duffey struck out just 21.7% of the batters he faced. With his new spin heavy approach he’s now struck out more than one-third of the batters he’s faced each of the past two seasons.
The season before Duffey made his adjustments, the club used a similar approach to turn reliever Ryan Pressly into a valuable trade chip. A former rule 5 pickup from Boston, Pressly did decent work in mostly low and medium-leverage spots for those terrible Twins teams 2013 to 2017. In 2018, Pressly made an adjustment that pushed his slider and curveball usage over 50% of his pitches, which, like Duffey, allowed him to get more swings and misses. It took him a bit to get comfortable with the new approach, but Pressly posted a 1.46 ERA in July of that season, striking out 30% of the batters he faced that month. With Minnesota on the playoff race fringe, the improved Pressly was traded to Houston for prospects Jorge Alcala and Gilberto Celestino at the trading deadline. Houston continued and even more heavily emphasized his new breaking ball approach. Pressly would allow only a single run in 21 innings down the stretch in a high-leverage role for the Astros, striking out 35.8% of the batters he faced in the second half of that season. With this approach, and additional tweaks from the Astros, Pressly would go on to set a Major League record with 40 consecutive scoreless appearances between 2018 and 2019.
The Twins have also made it a priority to acquire pitchers with this ability, often for minimal cost. In 2018, they signed journeyman right-hander Ryne Harper to a minor league deal. Harper would spend 2018 between AA Chattanooga and AAA Rochester, posting an 86-10 strikeout to walk ratio in 65 innings. Harper, a former 37th round draft choice, would make the 2019 Twins out of spring training and go on to appear in 61 games that season, giving a 3.81 ERA / 3.66 FIP in 54.1 innings. Harper was a valuable reliever (0.9 fWAR) despite well below average fastball velocity (89.3 mph, 7th percentile). The key for him was a big breaking curveball that he threw 60% of the time and manipulated it’s velocity and shape from the high 60s to the upper 70s. Collectively, hitters batted just .220 against Harper’s curveball and he generated a swing and a miss 30% of the time with the pitch.
With Harper’s 2019 team leading the AL Central by mid-season, Falvey looked to reinforce the bullpen at the trade deadline. Despite rumors of a big name acquisition, the Twins quietly added veteran right-hander Sergio Romo from Miami. Long a quality big league reliever and sometimes closer, Romo was already a heavy breaking ball user. Romo had often thrown his side-armed slider more than half his pitches throughout his career. In about a season and half with Minnesota, Romo would turn to his slider more than 60% of his pitches, giving the team 41.2 innings of 3.67 ERA / 3.80 FIP ball. Batters hit just .208 and struck out 27% of the time against Romo during his time in Minnesota despite fastball velocity that ranked in the 1st percentile (about 85 mph).
Continuing the theme, the team claimed to that point undistinguished reliever Matt Wisler off waivers from Seattle last offseason. Wisler was a former top starting prospect who had been a key part of a major trade earlier in his career. Over five MLB seasons before Minnesota claimed him, Wisler had never produced even an average season and his best season ERA was 4.28 over only 40 innings in 2018. In 2019, over 51.1 innings with San Diego and Seattle, Wisler pitched to a 5.61 ERA and allowed 1.75 home runs per nine innings pitched. His acquisition by Minnesota was seen as nothing more than a depth move, except for his ability to spin an effective slider. Hidden in his overall poor numbers, Wisler had thrown 70% sliders in 2019, holding batters to a .206 average and generating 40% whiffs with the pitch. The Twins had Wisler lean into that approach even more. Last season, his lone campaign with the Twins, Wisler threw an absurd 83% sliders. The approach worked and he had his best major league campaign, delivering 1.07 ERA / 3.35 FIP (both career best) and a 32.7% strikeout rate.
Finally, the Twins haven’t limited this approach to relievers. One of the two major offseason additions made to the 2020 team was the trade acquisition of starter Kenta Maeda. Frequently the odd man out of the starting rotation down the stretch in Los Angeles, the Twins deployed a similar plan to maximize Maeda’s value as a starting pitcher. In L.A., Maeda had thrown his four-seam fastball and slider each about a third of his offerings in 2019. With Minnesota, the slider became his primary offering (38.6%). Maeda decreased his four-seam fastball usage to just 18%, while also increasing his changeup usage. With the new off-speed heavy approach Maeda had the best season of his career with a 2.70 ERA / 3.00 FIP over a team-leading 66.2 innings. That production was worth 2.1 fWAR (2nd-most among AL pitchers) and also came with the highest strikeout and ground ball rates of his career. Hitters batted just .216 against Maeda’s slider. He would start game 1 of the 2020 AL Wild Card series against Houston and eventually finish second in the AL Cy Young voting.
With Duffey, Maeda, Romo, Wisler, and curveballing lefty Rich Hill (another off-season addition before 2020), the 2020 Twins threw baseball’s highest percentage of breaking pitches (40.6%). The team’s 28.0% slider rate was highest in baseball and it’s 12.6% curveball rate was 9th-most. As you might expect, the club’s fastball rate declined in tandem.
The 2017 Twins threw baseball’s fifth-most fastballs. The 2018 team threw the 4th-most. By 2019, that fraction decreased to just 19th-most as the new organizational philosophy took hold. Last season, Falvine’s Twins threw the 2nd-fewest fastballs in the game.
Not only did the team throw a bunch of breaking pitches, those pitches were highly successful. Collectively, the league hit .196 / .256 / .329 against Twins’ breaking pitches in 2020, good for a very weak .253 wOBA. Those numbers indicate Minnesota’s pitchers turned the entire league into Pedro Florimon (.254 wOBA in 682 PAs with Twins 2012-2014) with their breaking pitches. As measured by 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. The last time one team pulled that off was the 2009 St. Louis Cardinals.
What this has enabled is more strikeouts. In 2019, Minnesota pitchers averaged a strikeout per inning for the first time in franchise history, before improving on that again in 2020. They finished with an above average strikeout rate in both seasons.
After years spent wandering the pitching wilderness with little success or coherent direction, Minnesota’s new leadership team has successfully ridden modern approaches (and lots of spin) to turn pitching into a core strength of the organization and unlock major value for minimal additional cost.
But breaking balls are just the first part of the story, albeit a visible one. Pitching is a complex series of interactions between batter and pitcher. It’s a tricky proposition to point to any one thing as the full explanation for a pitching performance change. Minnesota’s performance the past two seasons is no exception. It’s too simplistic to just point at the Twins’ new reliance on breaking pitches as the reason for the team’s improved pitching performance.
The fact is, the focus on spin has also gotten the team better production out of its fastballs too. I’ll dig into that in part 2.
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.