The Twins can’t handle the heat

Through 55 games in 2020 the Twins have continued hitting bombs. They are on a pace that would equate to 250 homers over a normal 162 games – not the record pace of last year, but a figure that would easily rank as second most in franchise history. While the bombs are present, the overall offensive success has not been. At 4.45 runs scored per game, the Twins rank just 19^th and have posted a .242 / .316 / .429 (BA / OBP / SLG) composite line that is a far cry from last year. It’s not that the offense has been “bad,” after all, those numbers are middle of the pack-ish (101 wRC+). But they certainly aren’t what we expected and our expectations color how we evaluate and feel about the results.

There are a lot of possible factors contributing to the 2020 team’s lower offensive production. Injuries to key players and the likely related struggles against left-handed pitching are two obvious ones. General regression from a historic offensive peak is another, potentially more accurate, explanation. A fourth, which I want to dig deeper into today, is the team’s production against fastballs.

First, a quick note of explanation. Fangraphs lumps fastball types together in the data points I will use for this and the data counts four-seamers, two-seamers, and sinkers together as “fastballs”. Cutters are excluded.

★ ★ ★

The 2019 Twins earned the “Bomba Squad” moniker largely by obliterating fastballs. By Fangraphs’ pitch type linear weights – measures which answer the question “how well has a batter / pitcher (or team) performed against/using a certain pitch?” – last year’s Twins were the best collective group in baseball against fastballs (wFB). The Bomba Squad compiled 115.2 wFB in 2019, a figure that was well clear of the 2^nd-ranked Astros’ 99.2. That success against fastballs was the foundation to the team’s overall offensive success, as they posted a .270 / .338 / .494 composite line that included a record setting 307 homeruns and baseball’s second most runs scored (5.76 per game).

In 2020 the team’s collective wFB is 6.1 and ranks 14^th in MLB. For comparison, the league-leading Braves are at 47.3, with the Dodgers just behind at 45.6. wFB is a cumulative statistic, so comparing the raw totals to last season wouldn’t be fair. To adjust it for the games played disparity between seasons or players, Fangraphs provides an adjusted version of the stat, wFB/C, which is the weighted value per 100 pitches. This gives us a better way to compare apples to apples. By this, the 2020 club comes in at 0.16 wFB/C (still ranked 14^th). That’s decent and slightly above average but a significant decline from last season’s 0.93 wFB/C.

You might suspect a simple explanation is the league has adjusted to the Twins since last season and stopped throwing them as many fastballs to hit. The adjusted numbers above take care of that, but the raw pitch usage data also confirms that as a dead end.

Last year, Twins hitters saw the 2^nd-smallest proportion of fastballs in baseball, 50.5%. Only Baltimore saw fewer by Fangraphs tracking. As part of the multi-year game-wide trend of throwing fewer fastballs and more breaking balls, the 2020 Twins’ proportion of fastballs seen has decreased slightly to 49.3%. Interestingly, that figure is just the 10^th-fewest, despite the decline from last year. The decrease in fastballs thrown at the Twins is mostly insignificant. By Statcast’s pitch tracking, if the rate of fastballs thrown against Minnesota had remained constant from last year we could expect Twins hitters to have seen only about 33 more fastballs over the course of this season – not even an additional one per game.

If opponents aren’t using their fastball less against the Twins, perhaps another explanation could be that the fastballs they have seen have more velocity. Higher velocity pitches tend to be harder to hit and produce worse results. That’s a dead end too. The 2019 team faced fastballs that averaged 93.3 miles per hour. In 2020, it’s 93.2. Essentially the same. While using average velocity strips out variability that might be informative, it is a simple measure we can use to say the fastballs the Twins have faced haven’t been faster.

At the team level there doesn’t seem to be a clear factor we can point to that explains the decreased production against fastballs. We’ll need to dig deeper on the individual players to see if we can find something.

Here is a player comparison of weighted fastball production per 100 pitches, 2019 to 2020:

Only Nelson Cruz and Jake Cave have improved their production against fastballs from last year. Miguel Sano remains a solid performer in this area, but he’s come down slightly too. Same with Max Kepler to a larger extent. The 2020 newcomers aren’t listed in the table but have done their part to try to offset the decline of their teammates. Josh Donaldson is at 2.54, Brent Rooker was 1.55 in his limited appearances, and Ryan Jeffers is 0.59.

It’s clear the data in the table tracks well with our perceptions of the player’s performance this season, maybe save for Byron Buxton. Among the “regular” players, Mitch Garver, Marwin Gonzalez, Jorge Polanco, Eddie Rosario, Luis Arraez, and Max Kepler have all been the subjects of our fan disappointment and frustration at one point or another this season. Each has performed below our expectations in aggregate and against fastballs.

Let’s dig in deeper on the three regular players with the largest change (Garver, Gonzalez, and Polanco) and see if we can figure out what’s behind this for them.

★ ★ ★

Mitch Garver

wFB/C change from 2019: -5.62

Overall: With an injury wiping out a good chunk of the middle of his season, he’s been limited to just 70 PAs in 2020. It’s hardly enough to make any sweeping long-term conclusions, but those 70 PAs have been objectively terrible. His composite .148 / .243 / .197 line is easily a career worst and dramatically different than last year’s .273 / .365 / .630. After popping 31 homers (1 every 11.6 PAs) last year, Garver has just 1 homer in 2020.

Against Fastballs: After hitting .341 and slugging .838 against the fastball pitch group last year, pitchers have surprisingly continued to challenge him with fastballs in 2020. Garver has seen a slightly larger proportion of fastballs in 2020 (56.9% up from 55.9%). Despite that, in 39 PAs ending against a fastball Garver has hit just .152 and slugged .242. His Statcast expected batting average is .191 and expected slugging is .305, confirming that this is not a bad batted ball luck story.

Changes: While small sample size caveats definitely apply, it looks like there are three factors at play for Garver. The biggest change has been his ability to make contact against heaters. Garver’s whiff rate has doubled against fastballs, up to 32.5% after just 15.8% last season. Those swings and misses explain why his expected stats in the paragraph above are so low even though his average exit velocity on fastballs remains a strong 93.1 mph. When he hits it, he’s still hitting it hard. His overall swing rate is mostly in line with last year and he’s made improvements by swinging more often at pitches in the strike zone and decreasing his chase rate. He’s just not making contact nearly as often.

Another change is the profile of those batted balls. More of them have been pulled on the ground. Garver’s groundball rate against fastballs has spiked to 50% (last year 32%) and his pull rate is up to 59.4% (47.8% last year). Groundballs turn into outs more than other batted ball types and they don’t go for extra bases nearly as often. Opposing teams have responded to this by shifting their infielders against Garver in almost half his plate appearances (32 of 70), after doing so only 1 out of every 5 PAs last season.

Conclusion: Swinging and missing at pitches in the zone and pulling ground balls are signs of a hitter who has not found his timing. Given the disruptions of the 2020 season, a short summer camp, and a lengthy injury absence, it’s understandable how timing would be off, especially for a catcher who has many other responsibilities. Decisions will need to be made about how much playing time he should get while he’s fighting his swing this badly in 2020, but I don’t think it makes sense to make any long-term conclusions from 70 PAs in 2020. The 2020 Garver is no more likely to be the real Mitch Garver than the 2019 Garver was.

Marwin Gonzalez

wFB/C change from 2019: -2.77

Overall: A career .261 / .317 / .413 hitter as a semi-regular player in Houston and Minnesota, Gonzalez has hit just .209 / .283 / .323 in 180 PAs in 48 games in 2020. Counter intuitively, Gonzalez has increased his walk rate to 8.9% (last year: 6.7%) and held his strikeout rate flat (20.6%) with his career average (20.2%). Gonzalez’s more granular plate discipline stats remain mostly in line with his career and last year. He’s chasing less out of the zone and making contact when he does swing at a similar rate.

A switch-hitter, Gonzalez’s batting average and on-base percentage are similar from both sides of the plate in 2020. He has hit a homerun every 36 plate appearances in 2020, which is in line with his career performance of one every 35 PAs. While the homeruns have remained similar overall, his isolated power (slugging minus batting average) is just .114, his lowest mark in the past 6 seasons. That number is driven in part by his right-handed power vanishing (ISO: .036 as RHB vs .156 career). All 5 of Marwin’s 2020 homeruns have come as a left-handed batter.

Against Fastballs: After hitting .307 and slugging .453 against the fastball pitch group last year, Gonzalez has cratered to .190 and .286 in 2020. Pitchers have thrown him slightly fewer fastballs overall (48.1%, down from 50.8%), but not dramatically so. As a left-handed hitter those numbers drop to .177 and .274. Batting right-handed it’s .227 and .318.

Changes: There appears to be two factors at play for Gonzalez in 2020. The first is simply batted ball luck, especially as a left-handed hitter. His overall .233 batting average on balls in play (BABIP) is a career-worst and well below the .291 MLB average. His BABIP is just .213 when Gonzalez bats left-handed. Statcast calculated expected metrics indicate that Gonzalez should have a .242 batting average and .461 slugging percentage based on his batted ball profile as a left-handed hitter – numbers that would be much more palatable than his actual results. They are dragged down a bit by a career-high 17.8% whiff rate against fastballs as a left-handed hitter, but his overall batted ball profile (i.e., rates of groundballs, flyballs, and line drives) remains similar.

In a smaller sample size as a right-handed hitter (72 PA), Gonzalez has made more contact and hit the ball harder. His average exit velocity as a right-handed batter against fastballs is 92.7 mph, a career-best against that pitch group. Despite the increased exit velocity, his expected slugging is only .347 – significantly lower than last season’s .478 mark. The culprit here appears to be a change in his batted ball profile that is yielding more flyballs. Gonzalez’s average launch angle from this side against fastballs has spiked to 24 degrees, creating a three-fold increase in his fly ball rate (38.1%; last year: 12.3%). While flyballs are generally a good thing, hitting them too high can create easy flyouts.

Conclusion: As a left-handed hitter it seems a lot of Marwin’s struggles can be chalked up to poor luck and a few more swings and misses, both things that might correct themselves in a larger sample. As a right-handed batter he’s hitting the ball harder and in the air more than ever. Too often these batted balls have been on the trajectory for can of corn flyouts instead of productive doubles and homeruns. This helps explain his anemic ISO mark as a right-handed batter and how Gonzalez went from hitting a double every 20 PAs in his career, to one every 60 PAs in 2020. It seems there is a mechanical adjustment needed with Marwin’s right-handed swing.

Jorge Polanco

wFB/C change from 2019: -2.10

Overall: Another switch hitter, Polanco’s .264 / .313 / .363 line is 15% below league average in 2020. Most concerning is the sudden lack of power. The slugging mark is easily a career worst in a full-season for the shortstop. Unlike Gonzalez, Polanco’s issues have mostly been from the left-side of the plate, where his .230 / .296 / .324 line has generated only a 72 wRC+. From the right side, he’s hit .352 / .357 / .463 (122 wRC+). These marks are completely inverted from 2019, when Polanco was far better as a left-handed hitter than right (133 wRC+ as LHB vs. 88 wRC+ as RHB) and his career numbers (112 wRC+ as LHB vs. 89 wRC+ as RHB).

Against Fastballs: Polanco is a bit of a confounding case because his top line numbers against fastballs don’t look that bad – batting .306 and slugging .411. Those are generally consistent from both sides of the plate and he’s swinging and missing against fastballs at a lower rate than last season (8.4% vs. 11.7 last year). Context is important, though, and slugging .411 is a major decrease from the prior two seasons when Polanco slugged .527 and .526 against fastballs.

Changes: Focusing on the left-side of the plate, there are two changes that I believe contribute to Polanco’s lost power – an increased groundball rate (and lower launch angle) coupled with a significantly lower average exit velocity. Against fastballs from the left side, Polanco has hit the ball on the ground 36.7%, up from last season’s 22.9%. In conjunction, his average exit velocity has dropped to a career low 85.8 mph against fastballs (last year 91.2 mph).

Conclusion: Combined, slower hit ground balls turn into more outs, which is evidenced by a BABIP of .264 from the left-side (career .317). Those slow groundballs also yield fewer extra base hits which helps to explain the missing power. It’s not all a bad story, though. Polanco’s Statcast calculated expected slugging comes in at .491 overall and .490 from the left side. While not quite up to last year’s banner year, those figures would show favorably with his career marks. Perhaps he’s not that far away from his usual self.

★ ★ ★

Wrap Up

As a group, the Twins hitters have performed significantly worse against fastballs in 2020. That lower production has been driven in large part by Mitch Garver, Marwin Gonzalez, and Jorge Polanco. Some regression from last year’s peak was expected, especially for Garver and Polanco. In the small sample size that is 2020 Major League Baseball, it’s tough to know for sure if any of these changes represent a permanent shift in their baseline talent level instead of just the normal variation that we should expect to occur in any season. At any rate, for these three, it seems there are some reasonable explanations for their struggles against fastballs. Perhaps with some slight adjustments, Garver, Gonzalez, and Polanco can resemble their former selves sooner rather than later.

---

John is a contributor to Twinkie Town with an emphasis on analytics. He is a lifelong Twins fan and former college pitcher.