clock menu more-arrow no yes

Filed under:

An Update on the Cards’ Expected vs Real Batting Stats

New, 10 comments

Looking back at an earlier post, and checking to see what’s changed since then.

Cleveland Indians v St Louis Cardinals Photo by Dilip Vishwanat/Getty Images

A while back, I tried to examine the Cardinal offense through the lens of Baseball Savant’s expected batting stats, i.e. that newfangled Statcast stuff you’ve been hearing so much about in the celebrity rags.

I expect that most of you who frequent this site and others like it, dedicated as they are to hardcore analysis of the game, are familiar with the concept of xwOBA by now. If you’re not, a quick primer: Statcast measures basically everything, as best as we can possibly account for “everything”. Launch angle of batted balls. Exit velocity of batted balls. The percentage chance that a ball hit at a specific velocity and launch angle usually turns into a hit, or an extra-base hit. And then the whole kit and kaboodle is fed into a giant algorithm that spits out an expected batting line. What we’re trying to measure here, you understand, is quality of contact, and how a batter ‘should’ be performing given the kind of contact he’s making, combined with his plate discipline profile. Got it? Good.

So that was one of the lenses through which I attempted to examine the Cards’ early-season offense, which had been incredibly frustrating, to see if there was much reason for optimism. The good news: it looked at the time like the Redbirds, as a club, were hitting the ball better than the results had been. The bad news: the offense hasn’t gotten any less frustrating since that time, with a few notable exceptions. (Suck it, Kluber! Actually, you’re one of my favourite non-Cardinal pitchers in all baseball, so no offense, okay? But still, suck it.) At no point in time this season has this offense ever really clicked for anything more than a game or two at a time, and usually not more than one or two hitters in any given stretch.

There was some discussion at the time of how predictive xwOBA was toward future performance, and most of the research done so far seems to suggest it’s really more descriptive than predictive. As in, xwOBA very much captures what we should have expected in the past, and does so really well. Predicting future performance, though? That’s less certain, as the fact a hitter has hit the ball hard up through a given date is not a guarantee he’ll continue hitting the ball hard. Contact quality is noisier than something like plate approach. All that said, you certainly prefer to look at a player and see he’s hitting the ball hard and getting unlucky, rather than vice versa. There is some predictive aspect to contact quality, just not extremely strong predictive power.

So let’s revisit the Cardinal offense and see how things have shaken out since early May, shall we?

When I did this before, there was an interesting gap to be observed in the xwOBA vs real wOBA numbers, league-wide. As a whole, baseball was underperforming the xwOBA numbers by 23 points; i.e. a guy whose contact and plate discipline would suggest he should be a .350 wOBA hitter was, on average, a .327 wOBA guy on the seventh of May. Now, we don’t have all that long a history with these numbers, so it’s possible xwOBA always sees a little too high a baseline league-wide, with the falloff perhaps being defensively oriented, or maybe this year was especially bad early on with the strangely frigid April (‘Strangely frigid April’ sounds like a line from some TMNT fanfiction, by the way), we saw this year. Personally, I think the weather had a lot to do with it, and maybe that’s a yearly thing as well that was perhaps exceptionally rough this year.

A point in favour of my observation/belief: the league-wide gap in xwOBA-wOBA is down to .015, with there still being a decent shortfall in expected slugging, but expected batting average is only five points higher than the actual batting average for all of baseball. That would seem to make sense to me, really. It’s worth keeping that in mind, though, as we look at some Cardinal numbers; that throughout baseball, we see a fifteen-point deficit, so any player with a .015 shortfall or less is pretty much right in line with the game as a whole. If xwOBA is a sound measurement, we would expect to see it converge with the actual numbers, and that’s basically what we’re seeing.

So what about the Cardinals? Well, here’s some good news: if we look at the league-wide xwOBA leaderboard (seeing the minimum PAs to 200 in order to try and capture only full-time players), we find that Jose Martinez and Matt Carpenter rank eleventh and twelfth, respectively, in xwOBA. For reference, the only other teams who rate two players within the top 20 by this measure are Boston, with three (!), Cincinnati, with both Joey Votto and Eugenio Suarez in the top fifteen, and the Yankees, with Aaron Judge at fourteen and Gleyber Torres just sneaking in at 20. So going by plate discipline and quality of contact, the Cardinals have had two of the dozen best hitters in baseball in the lineup this year.

Now the not as good news: if we look at just the Cardinal hitters, we see some very good performances, but we also still see big gaps between the actual numbers and what the expected numbers tell us should be happening. The Redbirds’ five best hitters are, in order: Jose Martinez, Matt Carpenter, Yadier Molina, Marcell Ozuna, and Tommy Pham. Those five hitters are all in the negatives as far as xwOBA - actual wOBA. The deficits are as follows: -.039, -.042, -.048, -.040, -.041. In other words, the best hitters on the Cardinals are all missing approximately 40 points of wOBA or so this season.

We can see the shortfall is primarily power related; Jose Martinez is slugging 74 points below his xSLG, Matt Carpenter is off by 82 points, Yadi is missing 56 percentage points, Ozuna is under by a shocking 93 points, and Tommy Pham ‘should’ be slugging 86 points above what he actually is this year. And remember, these numbers take into account both exit velocity and launch angle, so it isn’t as if Statcast just thinks Marcell Ozuna is hitting the ball really hard and doesn’t realise he’s pounding grounders night after night.

There are a few overperformers, to be fair; Jedd Gyorko, Yairo Munoz, and Harrison Bader have all enjoyed some good fortune. Bader I’m not surprised by; he’s gotten a fair amount of value from his wheels this season, and speed is one thing not baked in to these expected metrics. Munoz runs well also, so maybe that explains some of his good fortune. Jedd Gyorko has been bad this year, but he actually probably should be horrible, rather than just bad.

Oh, and in case you’re wondering, Dexter Fowler has been unlucky, but also bad even by expected stats. And Kolten Wong has been bad, and somewhat lucky to rise to even that level. Shit’s rough, man.

Now, the thornier question to contend with here is why so many Cardinal hitters appear to be underperforming so badly, and to be honest, I don’t really have an answer. We don’t have all that many years of this kind of data, and that means we don’t really have a great feel for what kinds of things could cause systemic underperformance. Is Busch Stadium such a hard place to hit that most of a regular lineup could struggle to turn results into stat-sheet results all together? No clue. Has the weather been exceptionally bad for the Redbirds? Doesn’t really seem that way to me, but atmospheric science is also not exactly my forte.

Somewhat funny story: my girlfriend yesterday evening was texting me about the incoming storm, wanting to know if I had put Rags (short for Ragamuffin), my outside cat, in the basement in case a tornado came through. I had, for the record, but the next thing I received from her was this: “No, it’s not a supercell, it’s a linear storm with an inflow notch.” And then: “Shit. LOL. That wasn’t for you.” She had meant that for her best friend, who is also a huge weather nerd, apparently. I responded: “I’m guessing that’s how you feel when trying to read my baseball analysis?” She’s tried to read this stuff, but the acronyms and numbers are completely foreign. Sort of like an ‘inflow notch’ is to me. Unless it’s a euphemism I’ve just missed out on all these years, that is. Atmospheric science riff complete.

So in the end, I don’t really have any concrete conclusions about the Cardinal offense and why it appears to be underperforming so notably this year. The somewhat good news is that the gaps between expected and actual performance have narrowed somewhat since I wrote about this in May, and for the most part that has involved more improvement of the actual numbers than cratering of the expected numbers (Jedd excepted, sadly), but there are still big gaps in production for the majority of the Cards’ most important, and productive, hitters. We would probably feel a whole lot differently about this club and this offense if a couple hitters were slugging nearly a hundred points higher than they actually are.

Thus, I offer this more as a record of a moment in time, and a consideration of what the expected numbers seem to suggest about the Cardinal offense, rather than a diagnostic or explanatory post as to the why of the thing. I’ll revisit this topic again a month or two down the road to see where we’re headed. Let’s all keep our fingers crossed that the real numbers keep going up more than the expected numbers go down.