The Link Between Travis d’Arnaud’s Set-Up and Struggles

In 2015, Travis d’Arnaud was one of the league’s best power hitters. His .218 ISO placed him in the neighborhood of sluggers like Joey Votto and Kris Bryant. Following the season, Steamer projected that d’Arnaud’s ISO would be fourth best among catchers, and 24% better than 2015’s league average.

But that power was absent this past year, as d’Arnaud’s ISO fell by two thirds. At .076, it was one of MLB’s worst 10 marks, ranking the Mets catcher amid weak-hitting middle infielders like Dee Gordon, Adeiny Hechavarria, and Ketel Marte. d’Arnaud’s overall output took a huge hit, as his wRC+ sank to 74 this year after reaching 130 in 2015. That 56-point plummet is among the 1.1% worst year-to-year differentials of all time (minimum 250 PA). A decline this severe is unusual — and particularly surprising for a player who looked like a burgeoning star in 2015.

How did this downturn happen? In other cases, we might point to injuries or small sample sizes, but there’s reason to think that more was at play for d’Arnaud in 2016. That’s because he struggled with a longer swing in the 2016 season, generated by his bat wrap.

To wrap the bat generally means this: the hitter points the bat head towards the pitcher to elongate the swing path. Basically, the bat takes a longer journey before it enters the hitting zone. Seldom do baseball people speak admiringly of long swings, and wrapping tends to have a negative connotation. But it’s wrong to say that a wrap is inherently bad for hitting, because plenty of major leaguers wrap the bat (such as Miguel Cabrera, Bryce Harper, and Andrew McCutchen) to generate power and enjoy success.

Notice in the link above that these three hitters stick the bat behind their own heads. d’Arnaud’s bat wraps are different—his bat gets pointed towards the pitcher, but from in front of his body. You can see that on this swing from August 8, 2015, a home-run cut on a Jeff Locke pitch.

d’Arnaud’s bat starts out tilted slightly towards the pitcher and becomes more angled as he steps forward. But he doesn’t stay wrapped for long, and unleashes a strong, fluid swing. Compare that side view with this next one from Game One of the World Series two months later.

In this clip, d’Arnaud sets up with his hands wrapped and hanging over the plate. Midway through the pitch’s flight, he gets deeper into the wrap, as the bat comes even closer to parallel with the ground. This wrap creates a longer, loopier bat path, one that leaves less margin for error and disrupts the hitter’s timing. So you can see why these big wraps could present issues for d’Arnaud. But to figure out what role these set-ups played in his uneven performance across the past two years, we need to investigate more thoroughly.

First, we need to distinguish when d’Arnaud hit with the lightly-wrapped and the deeply-wrapped set-ups. To that end, I watched every 2015–2016 d’Arnaud batting video stored in the archives of Baseball Savant, and took a screenshot of every stance. For consistency, I grabbed snapshots from the very beginning of each pitch’s journey: as the ball left the pitcher’s hand (or as close to this point as possible). After reviewing those 98 images, I was able to identify five distinct periods of form in his stance, which are captured in the videos below. d’Arnaud’s mechanics aren’t a video-game-like picture of consistency within these periods, but after viewing just a few of the screenshots, you’ll see a tendency in whether his hands are or aren’t in that extended, deep wrap when the pitcher releases the ball.

In Period No. 1, spanning the start of the 2015 season through late August, d’Arnaud holds the bat either in an upright position or a minor wrap. His swing stays compact throughout the time frame.

Even though d’Arnaud’s 2015 season had progressed well through the end of the summer, he made a lasting change to his set-up. We can see that from August 24, 2015, and onward, d’Arnaud’s bat hangs over the plate in the deeper, extended wrap just before the pitch is delivered. Period No. 2 runs through the end of the season.

In period #3, d’Arnaud’s 2016 season begins with more traditional set-ups, if inconsistent ones; the bat is held behind his head early in April, and resumes its upright position later in the month.

The next time frame spans from the end of April 2016 through July. In Period No. 4, d’Arnaud’s bat is deeply wrapped and drooping over the plate.

Period No. 5 begins at the start of August. There are a couple of exceptions and the bat does move around, but d’Arnaud’s wraps are toned down and he takes shorter swings to finish out the 2016 season.

We can see differences in d’Arnaud’s set-ups, but what do they actually mean for his offensive output? Injuries have sapped him of plate appearances in the past two seasons, limiting the explanatory power of wOBA. But exit velocity can guide us here. Even when armed with small subsets of balls in play, a player’s “true” exit velocity quickly shines through.

So I sorted d’Arnaud’s 2015–2016 batted balls into the five buckets, and found his average exit velocity in each. I also broke down the rate at which d’Arnaud pulled balls in each period, as Pull% also “stabilizes” rapidly and gives us a window into whether d’Arnaud profiled as a different hitter in the five periods. For good measure, both statistics are adjusted for opposing pitcher tendency and handedness all throughout this piece.

Within the 2015 season, d’Arnaud’s exit velocity dipped by 0.3 mph when he went from the minor wrap (Period No. 1) to the deep wrap (Period No. 2). The differences were far greater within the 2016 season, when his batted-ball speeds dropped 2.1 mph from lesser-wrap Period No. 3 to deeper-wrap Period No. 4, and crept back up by 1.4 mph when the wrap waned in Period No. 5.

Those hit speed differences are sizable: one fewer mile per hour in average exit velocity corresponds to 7.6 fewer points of wOBA, and those two stats are highly correlated (0.54, at a 300 PA minimum). Gaining or losing a couple of mph in exit velocity indicates significant changes in d’Arnaud’s offensive ability from period to period.

Of even greater note are the huge jumps in Pull%. From Periods No. 1 to 2, d’Arnaud’s Pull% goes from very high to very low. That metric increases again in the bat-upright Period No. 3, moves downward in the extended-wrap Period No. 4, and reaches its highest peak in the less-wrapped Period No. 5.

All told, the Pull% numbers suggest rather resoundingly that his bat is late in the wrap periods. And remember: power usually comes to the pull side, which is why the top of the Pull% leaderboard is populated by above-average power hitters. d’Arnaud’s lower exit velocities and pull rates in the heavily-wrapped periods make for an unproductive combination.

To make the sample sizes more robust, let’s further aggregate the data. Periods No. 1, 3, and 5 are grouped as one overall phase (in which his wrap is minor), and Periods No. 2 and 4 are combined as another phase (in which he hits from that deep, extended wrap).

The same theme is evident: d’Arnaud’s ability to pull the ball and hit it with authority is diminished from the extended wrap. He loses 1 full mph off his average exit velocity, and experiences a substantial 11.2-point drop in pull rate.

The wrap could also change how d’Arnaud handles pitch locations. Does he see uniform reductions throughout the zone in wrap phases, or do his location preferences change? To answer that question, I averaged d’Arnaud’s batted-ball velocities by the wrap phase and the strike-zone quadrant in which the pitch crossed the plate. From there, I took the wrap-to-no-wrap difference within each quadrant. Negative numbers indicate worse performance when d’Arnaud’s bat is tilted more towards the pitcher.

d’Arnaud’s biggest change occurs in the up-and-in quadrant, where his average exit velocity declines by 3 mph when wrapped. Moving beyond that zone, he loses about 1 mph up and away, while holding steady low and in. And low and away, d’Arnaud actually gains 2.7 mph. Hitters with longer swings are better suited to hit low pitches, so it’s not surprising to see that low-and-away increase. But better performance on low-and-away pitches likely will not allow d’Arnaud to regain his power in full, as it tends not to be an ISO-friendly zone for a righty. And, as you can see in the Pull% delta table below, d’Arnaud’s Pull% on those pitches didn’t change much, anyway.

The differences in wrap-to-no-wrap Pull% are even more glaring in the zone breakdown. D’Arnaud hits the ball to the left side a whopping one-quarter less of the time on up-and-in pitches. And despite no loss in batted-ball speed on low-and-in pitches, d’Arnaud is pulling the ball much less when wrapped. Small Pull% gains on away pitches aren’t enough to make up for the lost pull power. Altogether, these results come off as an indictment of d’Arnaud’s deeply-wrapped set-ups. Consistently, we see stronger exit velocities and pull rates when the wrap is reduced.

What is d’Arnaud’s outlook for 2017? Through the ups and downs of the past two years, Steamer looks into its crystal ball and projects… that d’Arnaud will be a league-average hitter with a league-average ISO. That will disappoint those who saw stardom for the longtime prospect, but there does seem to be an avenue in which this is an underestimate. Projection systems don’t see mechanical flaws, and if d’Arnaud can cut down on the bat wrap and more frequently put himself into a primed position to hit, he may be able to erase that massive performance drop-off from this year.