Matt Carpenter’s injury adjustment

No one needs me to say it, but I will anyway: Things are looking grim. After suffering a three game sweep at the hands of their biggest competition for the Central division, the Cards sit 6 games out of the division and 4 1/2 games out of the second Wild Card game.

That said, no one’s giving up. Matt Carpenter recently admitted to dealing with a shoulder injury. Games are still relevant though, and he’s trying to push on. Most MLB players are extremely competitive, and Matt Carpenter shows no signs of being any different. Everyone’s hurt at this stage of the year, it’s going to have to be something serious before a competitor like Matt Carpenter is willing to cede his lineup spot.

However, one thing sure seems to be true: Something is wrong with Matt Carpenter. It also seems like he knows it, and is consciously compensating for it. It’s a little subtle, but I think I have a body of evidence to show it at this point.

Here’s the first piece of evidence:

As the graph indicates, this is a rolling 15-game average of Carpenter’s career swing rates, provided by the good people at Fangraphs.com. For a more interactive version, which shows you the actual numbers when you hover over a spot on one of the lines, check here.

The O-Swing% rolling average comes at an all-time low at present, at just 10.7%. That’s even lower than any point in 2014, which is his lowest season rate. The lowest rolling overall swing% shown here still belongs to 2014, but he’s in the same neighborhood.

21 months ago, my very first article on this fine site was about Matt Carpenter, and how he seemed to constantly change his approach. 2014 was the year he deliberately took his patience to the extreme, posting the lowest O-Swing% and lowest Swing% in the league.

In 2015, he eased up a little in terms of patience, and traded some contact rate for power. That’s tended to be the type of player Carpenter has settled into: one built on power and patience. And that brings us to our second piece of evidence:

This is another 15-game rolling average, this time for Hard%, which is the percentage of his batted balls that were considered to be hit hard. While it took him a minute to get things worked out in 2015, since mid-2015 he’s enjoyed some pretty consistently good results in terms of hitting the ball hard. The huge drop off in 2016 coincides with his back injury. Again, for the interactive version, click here.

Things were looking really good in 2017. While he had some poor results to start the year, I made the case that Statcast saw a strong hitter that was getting unexpectedly bad results. Just a few weeks later, Matt started to see really good results, and the narrative was that he it was because he was moved back to the lead-off spot. Hard% agrees with Statcast though that he was just getting unlucky early in the year.

Matt’s Hard% fell off a cliff from July 18th on. He rebounded a bit in mid-August, but then fell again, leading to the need for the cortisone shot.

We can paint a similar picture using the Statcast-based metric xwOBA. xwOBA replaces the on-contact portion of wOBA with the league average production of that hitter’s assortment of batted balls in the Statcast Era (2015 and on). Here’s Carpenter’s xwOBA before and after July 18th, along with some other stats:

According to the actual results, Carpenter looks about the same over this time period. A little more walks, strikeouts, and hits on balls in play, and a little less power leading to almost the exact same overall value.

However, until July 18th, xwOBA saw a great hitter who was unlucky to only be running a 121 wRC+. Afterwards, xwOBA has seen a slightly below average hitter who has played up to a 122 wRC+.

To get a better look, we’ll break down Matt’s batted ball profile using Statcast’s six qualities of contact. Those are six different buckets a batted ball can be placed in, based on the Exit Velocity (speed off the bat) and Launch Angle (vertical angle off the bat). To best visualize those different groups, here’s Matt Carpenter’s radial chart from 2017:

The protractor-shaped image above is used to represent any batted ball in terms of Exit Velocity and Launch Angle. Each dot is one of Carpenter’s batted balls in 2017. The six shaded regions are the six qualities of contact.

For some context, here’s the league average production of each type of batted ball, along with the league average rate hitters hit each type and Carpenter’s percentages, again using July 17th as a border:

This paints a pretty clear picture. Carpenter held an above-average rate in all three above-average contact qualities. That he avoided topping the ball so often and held an above-average rate of “under” contact points to his approach towards hitting the ball in the air, where you have the best chance of doing damage.

Since July 18th though, he’s declined in all three good forms of contact, and increased his rates in the three below average batted ball types. Not only that, he’s been below average in all three good categories.

As the two rolling average graphs indicate, Carpenter didn’t adjust right away. He stopped hitting the ball hard, and then shorty after started lowering his swing rates. In the first table we see his walk and strikeout rate only marginally increased when using July 17th/18th as the border. Things change though, when you move it back to August 1st:

Here we see two very different approaches from Matt Carpenter. With the ability to calculate his non-contact wOBA (the portion of wOBA that comes from non-contact plate appearances: unintentional walks, hit by pitches, and strikeouts) and already having his xwOBA handy, I was able to work backward to find his xwOBA on-contact. For context, the average non-contact wOBA is .200, and the average on-contact wOBA is .370.

With the approach he started the year with, Carpenter was well above-average in non-contact situations. He was well above-average when making contact as well though, so there was no reason for him to maximize his walks like he did in 2014. Moving the cutoff to August 1st, Carpenter’s xwOBA improves, but there’s still a significant dip in production. His xwOBA on-contact however is just .327, well below the on-contact average of .370.

To counter, Carpenter is returning to 2014 form, at least in terms of plate discipline. He’s walked just one time less than he’s struck out since the beginning of August, leading to a non-contact wOBA that’s even larger than his on-contact xwOBA.

Carpenter can tell his batted-ball authority has dropped, even if the results are still very similar. He’s compensating by controlling the strike zone better than he ever has before.

Not only is working walks right now a better strategy for him, hopefully the lowered swing rate means less pain on the shoulder, and more strength on the pitches he does swing at. It’s hard to say though how much that saved pain to his right shoulder is cancelled out by him playing third, which he predominately has of late.

If you haven’t noticed, Jose Martinez has been hitting the crap out of the ball, and he’s been playing first, Carpenter’s position going into the year. Carpenter shifted to third which was recently vacated thanks to an injury to Jedd Gyorko. Gyorko is back on the roster now, but has been limited to pinch-hitting just once.

Carpenter is doing the best he can to deal with his injury. By xwOBA he’s still been above-average over his poor stretch. Hopefully the cortisone shot can help things, but I’m by not someone that can even pretend to know if it will help.

It’s probably a close call whether this compromised version of Carpenter is a better or worse option overall than Greg Garcia at third. Garcia offers better defense, but probably a weaker bat than even compromised Carpenter. Being that he’s a gamer and seems to be a leader in the clubhouse, I’m guessing he continues to play, albeit hopefully with more rest. Based on the stats, that’s probably not a bad decision. Just don’t expect him to do the same damage on-contact you’ve grown used to seeing the last two and a half years.