Cardinals have improved their defense, but by how much?

There are two main improvements that are driving the Cardinals second half success that they didn’t have in the first half. The first is a vastly improved bullpen, which is benefiting from a combination of addition by subtraction, better pitchers taking the more important spots, and some good old fashioned luck. The second is defense.

I wanted to see just how much better the team has been on defense in the second half as compared to the first. Now obviously, this isn’t a situation where I need to prove my conclusion. Nobody in their right mind would argue the first half defense was better. That is not the point of this post. Rather, I wanted to see just how much better it is.

Before I figure out what is different, it’s worth wondering what has remained exactly the same. Marcell Ozuna has played starting LF for both the first half and the second half. Kolten Wong has also manned 2B for most of the time, splitting with basically the same guys as well. Due to first half injuries, that’s pretty much it. Paul DeJong and Yadier Molina ended up missing enough time in the first half that the defensive numbers at their respective positions end up being affected. And now Ozuna is injured. Go figure.

First off, let’s compare first base from the first half to the second half. In the first half, Jose Martinez started 75 games while Matt Carpenter started 17 games. The script has virtually flipped. I already wrote a post about Jose Martinez’ defense so I will stick to what I concluded there. I’m putting him as a -12 defender at 1B, which is on par with the worst first base defender on record, Adam Dunn. Matt Carpenter has been exactly average at 1B over his career, but he’s played less than 2,000 innings when you really need 3,500. Considering he’s a -3.7 UZR/150 defender at 3B, he’s most likely better than average at 1B. Taking into account his 3B numbers, I’ll pencil him in as a +5 1B.

First half: -9.2 UZR/150

Second half: 0.3 UZR/150

Martinez has started 9 games at 1B in the 2nd half, so that’s dragging the numbers down, but you still see 9 extra runs saved on defense from the change in playing time. At 2B, the playing time in the first half is more split than you’d think. Kolten Wong takes 63 games, Jedd Gyorko 14, Carpenter 11, and Greg Garcia 6 games. In the 2nd half, Wong started 17 games, Yairo Munoz 11, and Garcia 7. I’m going to guess Carpenter playing at 2B even a little will make the 2nd half look better here. For what it’s worth, I am not taking Wong’s UZR numbers at face value, but combining his career and recent performance, I still have him as +8.9.

First half: 5.4 UZR/150

Second half: 4.4 UZR/150

I was wrong. Also if you were wondering, I pencilled Garcia, Gyorko, and Munoz all average at 2B. Gyorko has a good sample and is close enough to average anyway. Garcia and Munoz do not. Garcia has played about 500 innings at 2B, SS, and 3B. He’s been slightly below average at 2B and 3B and slightly above average at SS, which obviously makes no sense. Munoz has been absolutely dreadful by the numbers at each position he’s played in the infield, but he has such a ridiculously tiny sample, I can’t even pretend it’s worth looking at. Anyway, Wong played a higher portion of games in the first half so that explains the (tiny) disparity despite Carpenter’s numbers weighing the first half down.

At 3B, I feel comfortable giving Carpenter the -3.7 UZR/150 he’s had for his career - feels right and all. Gyorko has a +2.3 UZR/150. It seems like it could be higher but he doesn’t have the sample size. Even ignoring his innings in 2013, when you adjust for the sample size, it goes right back to +2.3. We’ve got some spot starts from Garcia and Munoz, but they are barely a factor since I’m still counting them as average fielders at the position.

First half: -1.5 UZR/150

Second half: +0.7 UZR/150

Onto shortstop, another position that would appear to be easy, but is not due to injuries. For the first half, Paul DeJong played only 47 games. Yairo Munoz played 32 games, Garcia played 10 and Gyorko played 4. Since I made Garcia and Munoz at 2B and 3B, it stands to reason, due to positional adjustments, to give them -5 each at shortstop. Having seen Munoz play SS, this seems generous, but let’s go with it. Paul DeJong, again accounting for regression due to a less than ideal sample size, ends up a +1.7 UZR/150 fielder at short. Let’s run the numbers.

First half: -1.6 UZR/150

Second half: 1.1 UZR/150

First base was obviously a HUGE jump, but the last two positions are very slowly adding to the total. Frankly, I think I’m underestimating the jump from Munoz to DeJong, but I’m trying to be conservative given the sample sizes. We can skip left field. In center field, the change has gone from Tommy Pham to Harrison Bader pretty much. In the first half, Pham started 79 games while Bader started 13 at center. I have Pham as a +1.1 UZR/150 fielder. Bader is a tad bit harder to try and guess for an exercise like this. Because he has a sample less than 1,000 innings, fielders will nearly always have something pretty close to 0.

So I came up with an imperfect solution. When you’re dealing with a small sample size, you need to fill in the blanks with average fielding. So for anybody who doesn’t have 3,500 innings or more at a position, you’re supposed to assume the innings he hasn’t played to get to 3,500 will be average. Instead of assuming the remaining innings played will be average, I’ll pencil him at a most likely too low +5 instead. So that’s my explanation for him being “merely” a +6.9 UZR/150 fielder (which by the way - still nice!)

First half: +1.9 UZR/150

Second half: +4.7 UZR/150

Lastly, we move to right field, which I’m very curious about given Dexter Fowler is not a particularly good fielder, but Bader played a not insubstantial amount of innings while he was here. Also Fowler played some innings in the 2nd half too, and so has Martinez. Since I concluded Bader is a +6.9 CF, that means he’s a +11.9 corner outfielder. Fowler is a -9.5 UZR/150 in his career in center, and using positional adjustments, that makes him a -4.5 fielder in RF. I don’t really know what to do with Martinez. I checked the worst UZR/150 of the past 10 years with a sufficient sample and Michael Cuddyer is the worst at -8.9. So we’ll go with that for Martinez, which seems basically fair.

First half: 0.3 UZR/150

Second half: -3.9 UZR/150

Lastly, I’m not going to attempt to figure out catcher defense, but I will attempt to capture the difference in not having Yadier Molina for a number of games. So don’t look at my guesstimates as an attempt at capturing anyone’s defense but an attempt to capture the difference between Yadier Molina and his replacements. So for simplicity’s sake, I’m assuming Molina as a +5 catcher. I am also assuming Carson Kelly is a +5 catcher. This may be harsh, but I’m assuming Francisco Pena is a -5 catcher. He just doesn’t seem like a very good defender. Classic backup catcher syndrome who everyone seemingly assumes is a good defender just because they are such bad hitters.

First half: +2.5

Second half: +4.1

It’s time to tally up the results and see how much better the Cardinals defense is using actual numbers. Combining all the positional changes, trades, injuries, the Cardinals come out to a team that is +13.6 runs better on defense than the first half. Now this doesn’t quite capture the “true talent” change in defense, because I counted injuries to DeJong and Molina. So let’s remove them. You still have a team that is +9.3 runs better on defense, because Wong is playing everyday, Martinez is in the less damaging RF, Carpenter’s at 1B instead of 3B and Gyorko is at 3B instead of Carpenter.

That is possibly too low though. Consider this. Wong got injured quickly after the All-Star break, being replaced by inferior defenders, which brought down the 2nd half numbers. Wong wasn’t injured in the first half. He just only played in ²⁄₃ of the games. If you compare Wong by himself with the cascade of 2B who played with Wong in the 1st half, you are 3.5 runs better.

Also, it’s easy to forget, but Bader basically didn’t start until Pham was traded. He was on the bench while the Cardinals were making moves to make sure he could get playing time. Now he’s playing everyday. Bader at +6.9 compared to the 1st half CFs (mostly Pham)? A 5 run upgrade by himself. Just by playing Wong and Bader everyday in the 2nd half, the Cardinals are saving 8.5 more runs in the 2nd half than the 1st, which is astounding when you consider both players still played in the first half and contributed somewhat to the first half numbers! (If you’re wondering, changing this puts the Cards up to +14 runs over the first half)

What is the most amazing about the defensive change is it hasn’t come at the expense of the offense. The Cardinals are both hitting better and playing better defense with this arrangement. Projection-wise, it’s basically Gyorko’s offense at 3B instead of Fowler in RF (since Martinez was at 1B, which meant Carp took Gyorko’s spot at 3B) and Bader’s offense instead of Pham. Offensively, Bader is a downgrade from Pham (though admittedly not when Pham literally isn’t playing because of an injury like now), but Gyorko is an upgrade over Fowler. The disparity between Fowler and Gyorko is probably less than the disparity between Bader and Pham, but I doubt it’s by a lot and it’s certainly well worth the very minor downgrade in offense in exchange for the defensive runs.