The Cardinals Have Been Shift Out of Luck

As of this writing, we’re approximately a fifth of the way in to the season and the Cardinals’ offense, on the aggregate, has been... fine. Entering play Tuesday, they were scoring 4.45 runs per game, just a tick below the league-wide average of 4.47. Their wOBA is 19th, OPS is 21st, and wRC+ is 17th as is ISO. Their plate discipline, as judged by BB/K, is a very healthy 4th. It’s a mediocre profile- neither spectacular or even good, nor terrible or even bad. There’s a glimmer of hope in there. Specifically, they’ve put up average numbers while three of their lineup anchors- Matt Carpenter (79 wRC+), Dexter Fowler (65), and Marcell Ozuna (71)- have been dreadful, approaching sub-Kozma levels. That they’ve done this with those three stalwarts stumbling is a testament to the early success of Tommy Pham, Paul DeJong, Jose Martinez, Jedd Gyorko, and some well-timed production from Yadier Molina and Kolten Wong. Even then, Wong himself is unburying himself from his own awful start, and Molina was fading before the dark times came. We won’t see him again for several weeks.

Lots of folks have (rightly) taken a peek at Statcast and discovered the giant gap between xwOBA (expected weighted on-base average) and actual xWOBA for many Cardinals. Indeed, if we look at xwOBA-wOBA for the 378 players with 30+ ABs thus far, Carpenter is 5th worst, Martinez is 14th worst, Molina is 23rd worst, Ozuna is 27th worst, and Fowler is 42nd worst. As we know thanks to great research by our own Tyler Kinzy, some of this can be explained by poor foot speed. That would certainly apply to Martinez, Molina, and Carpenter. It does not apply to Fowler and Ozuna, nor does it fully explain the gap for the other three. Add all of this up and you have a team that is dead last in xwOBA minus wOBA. No team has a bigger gap between their expected production, based on their batted ball profile, and their actual production:

It’s also worth noting here that every single team but two of them are underperforming their xWOBA, as A.E. Schafer/the Red Baron pointed out a few days ago. That’s extremely odd. But even within the context of a league with suppressed production relative to their expected production, the Cardinals have been very bad. There they are, way over to the right in that graph, worst in all of baseball.

According to conventional wisdom, if the team continues to hit the ball hard and at the various launch angles that have landed them in their current predicament, they will eventually be much more productive. I have no doubt that this is true. Where concern creeps in is the degree to which they will recover. Unfortunately, there’s another plausible explanation for some of the offensive struggles. You see, opponents are shifting where the Cardinals eat. And you never want anyone shifting where you eat.

To date, the Cardinals wOBA against the shift is .218. When opponents employ a shift, either traditional or non-traditional per Fangraphs’ definitions, 28 other teams (all but the Brewers) are more productive than the Cardinals. The good news is that they don’t particularly hit into a lot of shifts. As of Tuesday, their percentage of plate appearances against the shift was 27.7%, compared to the league average of 33.6%. The gap between their Shift-wOBA and Non-Shift-wOBA is very bad- .285 without the shift, .218 against it for a gap of .068, third worst in the league. This time, they still rank ahead of Milwaukee, and they barely leapfrog the Orioles. Their non-shift wOBA isn’t great, ranking only 22nd in MLB, but it’s still light years ahead of their production against the shift. The shift is killing them.

Let’s point some fingers. Below, you’ll see a tree map illustrating the percentage of team plate appearances against the shift that each player has taken, along with a simple bar chart showing each player’s wOBA against the shift. You’ll see some players missing- Molina and Bader, amongst others. That’s because these players have not faced a shift, according to Baseball Savant, entering Tuesday’s games.

Looking at the tree map alone, we can see that Carpenter and Fowler account for an enormous portion of the shifts the Cardinals have faced thus far. It amounts to 77.6% for the two combined (59.4% for Carpenter, 18.2% for Fowler). That’s not shocking. Also not shocking- all of Fowler’s PAs against the shift have come with him hitting left-handed. On the surface, those two would be the obvious culprits for why the team has poor numbers against the shift. But that’s not entirely true. Fowler’s production against the shift is just a little bit below league average (.280, compared to .290 league-wide). Carpenter’s at .272, which is below league average but still ranks third on the team. In both cases, they are well above the team’s collective .212 wOBA against the shift. They’re culprits for the team being below average, but they aren’t culprits for that dreadful .212 wOBA against the shift.

Now, let’s move on to the rest. It looks a little odd, but Ozuna and Wong have the same number of plate appearances against the shift. That’s why Ozuna’s bar is a little thinner than Wong’s. Cut me some slack- it’s the first tree map I’ve ever made. Before going further, we can free Tommy Pham from any sort of blame for the team’s struggles against the shift. Add “beating the ever-living piss out of the shift” to the copious amount of things that make us all love Tommy Pham. Sure, it’s a small sample (seven plate appearances), but it’s downright comical that Pham has been more productive against the shift than he has without any shift.

It’s that bottom group we need to evaluate. Munoz, DeJong, Garcia, and Martinez have collectively hit into the shift seven times and produced nothing. Nada. Zip. Zilch. After you throw in Gyorko’s .173 wOBA and Ozuna’s .211, you’re suddenly talking about 12.4% of the team’s shift plate appearances producing next to nothing. Wong’s .263 wOBA is better than these others, but it’s a little bit uncomfortably below average. Add Wong and now you’re at 18.2% of all shift plate appearances going to hitters struggling mightily with the shift. The shift has magically rendered about one-fifth of the team’s plate appearances against the shift as productive as plate appearances from Jaime Garcia or John Lackey. Ouch.

Thankfully, most of this can be dismissed as small sample size chicanery. Garcia, Martinez, and Munoz have only faced the shift once, and Munoz is now in AAA. DeJong and Gyorko each have four plate appearances against the shift. One double in the gap suddenly changes everything. I piled Wong into the “comparable to Lackey and Jaime as hitter” group, but he’s realistically been much better- closer to average than he is to the team’s dreadful .212 Shift-wOBA. Most importantly, he was actually well above league average at beating the shift last season. Unless something has drastically changed with Wong’s approach or unless the small samples for others are suddenly somehow predictable, we can absolve the team of most of the 18.2% of shift sins.

There are three remaining concerns, and it’s the usual suspects- Fowler, Ozuna, and Carpenter. Unfortunately, as amazing as Baseball Savant is, they don’t offer xwOBA against the shift. We can’t definitively link up the two sets of data- the gap between xwOBA and wOBA for Carpenter, Fowler, and Ozuna, and their respective struggles against the shift. There’s no hard proof that drastic shifts and their shift wOBA is what’s causing their actual wOBA to fall so much below their xwOBA. But I don’t think it’s an outlandish possibility. On the face of it, it sounds like a very plausible explanation. After all, if you’re hitting hard and not getting the results, it stands to reason that you’re hitting the ball at some well-placed defenders. And if it’s specifically shift-related rather than just dumb luck, that doesn’t bode well.

Simple dumb luck would imply that eventually those loud outs will start becoming loud hits. But if it’s part of a strict shift strategy against these hitters, the implication is much less optimistic. It means either the hitter will have to change their approach to beat the shift or the xwOBA-wOBA gap will continue to widen. For a guy like Ozuna, the percentage of plate appearances against the shift has been small, but his cratered 2018 plate discipline exposes him to a more pronounced effect. Fowler has been decent against the shift, and his struggles are more related to hitting right-handed (when he doesn’t face the shift). I think Fowler is a little insulated against shift shenanigans impacting his season.

And then there’s Carpenter. His production on shifts has been... reasonable. It could be so much better. The problem is that he’s facing the shift over 80% of the time. Done effectively, that’s going to take a huge bite out of his production. It’s especially troubling when combined with his stated revised approach this year- to back off of the flyball revolution a bit, and try to spray more line drives around the ballpark. The league has adjusted to Matt Carpenter. And yes, the effects are more pronounced than they should be. Some of that will wash out in the end. Sadly for us, some of the pronounced effect is by design. And for Carpenter and Ozuna, that dampens enthusiasm about the possibility for positive regression driven by xwOBA.