The Cardinals Offense: ZiPS vs. Reality

Somewhere around mid-May, the St. Louis Cardinals’ season completely derailed. Their performance since has been dreadful. Admittedly, there were warning signs before the season. Many projection systems saw the NL Central as a middling crapshoot between the Cardinals, Brewers, and Cubs, with the Reds standing on the periphery. Through the first 38 games, the Cardinals made those projections look silly, running out to one of MLB’s best records. Then everything fell apart, and here we are. They couldn’t salvage a win against the lowly Tigers. They lost at home to the Pirates. They haven’t even looked competitive and it’s an embarrassment. In fairness, the projection systems had them close to .500 entering the season, and that was before they saw IL trips for their starting shortstop, centerfielder, catcher, closer candidate, and two most important members of their rotation. Still, the most surprising and disappointing part of all of this has been the disappearance of the offense once the swoon started, injuries or not.

Since May 14th, the Cardinals have a non-pitcher wRC+ of 89, fourth worst in baseball. Their ISO is 26th in that time frame. Statcast tells a slightly different story (a respectable 14th best True Hard Hit% and 13th best True Barrel% since mid-May), but their paltry walk rate torpedoes the value. If you add True Hard Hit% and BB%, they’re 19th worst. It’s disappointing because they had made tangible strides early in the season at the plate. Given everything I’ve written thus far, the question is “Could we have seen this coming?” Well, the real question is “Why the hell did I spend part of my workday watching them lose to the @$%*ing Tigers on Wednesday?” but look... we all make bad choices sometimes. I digress.

As to whether or not we could have seen this coming, ZiPS is as good a place to start as any. There are a lot of projection systems out there, all with their own advantages, but I’m partial to ZiPS. Here’s a simple dumbbell graph of how Cardinal hitters have performed using wOBA through Tuesday’s game, plotted against their pre-season ZiPS projection. This is from the ZiPS DC version- adjusted for depth charts-projected playing time. I capped it at players with a minimum of 40 plate appearances. You’ll see Rondon in there, which was an accident when I made the graph since he’s under 40 PA.

As I clumsily tried to note in the key at the bottom, red bars are for hitters who are beating their projection while blue bars are for underperformers. Six of the 14 are beating their projections, while 8 are under. The quandary is that the projection-beating group is only modestly beating their projection with the massive exception of Tyler O’Neill and, to a lesser degree, Yadier Molina. Tommy Edman and Edmundo Sosa are almost exactly in line with their projections but technically qualify as ZiPS beaters.

On the flip side, there are some severe underperformances. Lane Thomas and Justin Williams have combined to give the team a ~200 plate appearance black hole in the outfield. Andrew Knizner’s audition in place of Molina was unproductive, well below his projected ZiPS wOBA. Through Tuesday, Paul DeJong was a full .050 below his ZiPS wOBA. Paul Goldschmidt has been respectable, but still below his ZiPS wOBA.

Some of this is at least a little irrelevant. Thomas, Williams, Knizner, Rondon, and Bader have been the biggest ZiPS underachievers, but all have limited playing time. With the exception of Bader, they’re all platoon players or bench contributors. Fewer plate appearances means more variance in their performance. A 3-for-4 with a couple of doubles from any one of them would dramatically shorten those blue lines in the graph. Which isn’t to say a 3-for-4 with a few doubles is likely for any of them. I only bring it up to demonstrate that their underperformance is hardly damning in any individual case. And a big ugly blue bar in the graph for Thomas (58 PA) should not have the same weight as the big ugly blue bar for DeJong (189 PA).

Taken on the whole, what does it mean? It’s hard to zero in on a specific wOBA for these players collectively since I don’t know the wOBA constants Dan Szymborski used for ZiPS. But we can spitball it. We’ll omit Rondon, as his inclusion in the graph was a mistake. With that adjustment, we get the following rough estimate:

ZiPS Projected: .311 wOBA
ZiPS Actual: .306 wOBA

No matter how they’ve arrived at that .306 wOBA from those players, it’s well within the ballpark of what you might have expected if you used ZiPS. If they were simply matching their projections, the team’s overall production might be a tick better- maybe they’d rank 20th or 21st in non-pitcher wOBA instead of 23rd- but their reality is pretty damn close to the projection.

This papers over a few things, of course. For starters, the team’s hope was that enough of these players would take a step forward to avoid having the 23rd “best” wOBA in baseball. O’Neill, Carlson, Arenado, and even Molina have compensated a little, with all four representing dramatic improvements over 2020. But it’s more than negated by dropoffs from the two Pauls, Bader’s absence, Carpenter’s inability to get even a dead cat bounce, and the complete lack of production from the Williams/Thomas/Knizner trio of offensive sterility.

Second, it also papers over the possibility that these players collectively could be outperforming their projections. There are error bars with these things, and I’ve fixated on the negative. A group collectively projected to have a .311 wOBA could reasonably be expected to run up an actual .306 wOBA. But they just as easily could reasonably be expected to run up a .316 wOBA, or more. That hasn’t happened.

Part of the problem is depth, or at least performance by the back end of the roster. If we look at non-pitchers with 75 plate appearances or less- essentially the platoon players, replacement-level players, and general bench chaff- the Cardinals’ wOBA is a paltry .208, 5th worst in baseball. Their wOBA from players with 75+ PAs isn’t exactly amazing- at .315, it’s tied for 18th- but their dropoff from “starters” to bench is 6th biggest in the game. I glossed over the poor contributions from Thomas, Rondon, and others but apparently it has been a drain on the offense. On the other hand, their percentage of plate appearances for bench contributors is 6.74%- ninth lowest in baseball. Put another way, the back of their roster has been awful, but at least it’s not playing much.

There’s ample room for potential causes here, and I’m sure fans have plenty of their own, but I’ll leave that for another day. In the very least, unfortunately, the team’s overall offensive performance shouldn’t be a surprise.