clock menu more-arrow no yes mobile

Filed under:

The Giants lead the majors in double plays turned. That’s good, right?

Here are 1,000 words in which I ponder the question, “Should we care about a contextual counting stat?”

Colorado Rockies v San Francisco Giants Photo by Thearon W. Henderson/Getty Images

You may or may not be aware that the Giants lead the major leagues in double plays turned. Not hit into. Turned! As in, they’re the ones doing the good thing. The 89 they’ve turned gives them a rather sizable lead on the next team, too. In Saturday’s game alone, they turned four double plays.

It should come as no surprise the Giants have turned so many double plays what with the slick fielding Brandon Crawford and Joe Panik up the middle with Kelby Tomlinson understudying them. Not to mention Ty Blach has thrown a large portion of their innings, and he has the third highest ground ball rate in the majors among pitchers who have thrown 50 innings.

Let’s appreciate some of the dope double plays the Giants have turned this year.

Yeah, nice.

That’s the stuff.

Hell yeah.

But what if, and bear with me here, what if turning a lot of double plays is actually bad? Like a few here and there is fine and good, but if there’s a lot it might be indicative of a problem.

To turn a lot of double plays, a team needs a ton of opportunities to turn do so. To have opportunities, a team needs to allow a bunch of baserunners. Does the fact that the Giants turn a lot of double plays reflect poorly on their pitching?

Down at the bottom of the double play leaderboard is the Houston Astros. The Astros lead the majors in ERA, FIP, xFIP. They also lead in WHIP, so they’ve been good at preventing baserunners. Makes sense then that they wouldn’t turn many double plays. They simply don’t have a lot of opportunities.

The Giants pitching staff hasn’t been anywhere close to as good as the Astros. The Giants have two pitchers in the NL’s bottom 25 pitchers by walks per nine: Chris Stratton and Derek Holland. Blach and Stratton are both in the bottom 15 by strikeouts per nine. The Giants have had their share of problems on the mound.

But they also haven’t been anywhere near as bad as the Royals. The Giants’ team ERA of 4.06 puts them at 15th in the majors, so smack dab in the middle. The Royals have turned a nice-not-great 69 double plays and have the worst team ERA in the majors. The Royal’s defense has only been about four runs worse by DRS, though admittedly there’s a 20-run difference between Brandon Crawford and Alcides Escobar.

So what is it? Does turning a bunch of double plays indicate that the defense (particularly the middle infielders) is good? Or does it actually indicate that the pitching staff is great (at giving up baserunners)?

Before I go any further, I think it will be helpful to go over two things. The first of which is correlation coefficient or r. This value is a number between -1 and 1. An r of 1 suggests a perfect correlation. For instance, how old you are compared to how many birthdays you’ve had would have an r of 1. Values between 0.2 and 0.5 suggest a weak to moderate correlation and values over 0.5 suggest a strong correlation. A value of 0 suggests no correlation at all. For instance, how many movies Nicholas Cage has been in compared to how many cats I’ve met would have an r of 0.

Second, these are broad looks across all situations. This doesn’t look at specific situations. It’s meant to be a cursory look and not exhaustive research. I’ll leave that to the people who didn’t major in English literature. The idea is to look at what a team that turns a lot of double plays is good at (or not good at), if anything.

If we look at all the team this year, there’s some evidence to suggest that teams who turn two haven’t pitched well.

That’s an r of 0.43, so that’s a moderate correlation. This is also only looking at a little less than half a season’s worth of games. If we look at last year’s scatter chart, the correlation diminishes.

That’s an r of just 0.19, so it’s a positive correlation but a weak one. My suspicion is at least partially correct, but turning a lot of double plays doesn’t appear to indicate that the pitching staff is terrible. It also doesn’t suggest that it’s helping the staff a whole lot. Otherwise, we’d see a negative correlation (the more double plays the lower the ERA).

Looking at 2017 numbers here are a few other correlation coefficients that I thought might be related to twin killings.

Pitching and Double Plays

Pair Correlation Coefficient
Pair Correlation Coefficient
DP and Baserunners 0.231
DP and GB% -0.033
DP and RE24 -0.097
DP and WHIP 0.230

There’s a positive relationship between double plays and raw number of baserunners (hits – home runs + walks + HBP), but it’s weak. Surprisingly, there’s no relation between double plays and ground ball rates, and if there is one it’s a negative correlation. One would think inducing more ground balls would help get more double plays, but that doesn’t seem to be the case. There also isn’t a relationship between double plays and RE24 (run expectancy over 24 out-states). I would think double plays would impact that, but I suppose run expectancy tables accurately predict the possibility of double plays.

So double plays don’t tell us a whole lot about pitchers. Teams that turn a lot of double plays tend to have slightly worse pitching staffs but not enough to worry about. Does it tell us anything about a team’s defense?

It does not. Of course, the majority of double plays are turned by the middle infielders. Does the defensive ability of a team’s second baseman and shortstop predict how many double plays might be turned?

It’s a stronger correlation, but not anything to care about. Its r of 0.20 is about the same as the relation to ERA.

Double plays then are a bit like RBI. They have virtually no predictive power and even predictive stats have trouble predicting them. Not only that, their descriptive power is limited at best. They can tell you who provided the excitement (or frustration) in a game, but over a season it probably doesn’t matter if the Giants turn the most or the fewest. But hey, at least it’s not a bad thing.