More Plate Coverage Isn’t Necessarily Better
Corey Dickerson is one of the best hitters at covering the plate. (via Arturo Pardavila III)
Writers, players, and managers alike often invoke the idea of “plate coverage” as a holy grail. See for example Mike Axisa’s take in 2016 on how Trevor Story’s plate coverage fueled his impressive start. Here’s one from Dan Connolly in 2015 about the then-hot Jimmy Paredes and his impressive plate coverage. And in 2014, Harold Reynolds and Dan Plesac produced a whole video on the issue for MLB Tonight.
Given these articles and similar widely held beliefs about plate coverage, you could be forgiven for thinking more plate coverage is better. And why wouldn’t it be? A hitter who can reach more pitches should be a more dangerous hitter, because they have more opportunities for contact.
But I was curious. Is this conventional wisdom true?
Methodology
To find out if more plate coverage is indeed better, I:
- Examined Gameday data for all balls hit fair from 2015 – 2017 by non-pitchers with at least 50 such balls hit. I didn’t count foul balls because to me, a foul ball doesn’t achieve the goal of plate coverage, which is to put the ball in play. I also didn’t count swings; those represent attempted plate coverage. I wanted to focus on actual plate coverage.
- Defined horizontal coverage as the median absolute deviation (MAD) along the x-axis and vertical coverage as the MAD along the z-axis (which is what Gameday defines as pitch height off the ground). Statcast gives this data in feet but for this article I’ve converted them to inches, because baseball is a game of them.
- Computed plate coverage in square inches by multiplying both MADs together.
Example: Plate Coverage in 2016
The following graphs show the worst and best horizontal plate coverages of 2016:
Delino DeShields can really cover pitches inside. Here’s one of them he got a handle on:
It’s a bunt, yes, but it’s still impressive coverage. But although he handled inside pitches well, the distribution of balls he was able to hit fair, along the x-axis anyway, was only 4.5 inches. That’s half the horizontal coverage of Caleb Joseph, whose distribution was twice as wide: 9.1 inches. The boxes in the boxplots tell the tale; DeShields’ box is much narrower than Joseph’s.
How about vertical plate coverage from 2016?
Corey Dickerson’s vertical coverage stretched from 0.8 feet off the ground to 4.5 feet off it, whereas Taylor Motter’s coverage extended from about 1.5 feet off the ground to 3.5 feet off it.
Combining both horizontal and vertical coverage gives us total plate coverage, in square inches. The leaders in this metric for 2016 were:
Joseph and Dickerson lead the list, as does free-swinger Salvador Perez. (Former) teammates Matt Wieters and Adam Jones, another free-swinger, come in next. Except for Joseph, these guys are all full-time players. Their bats can reach a large variety of pitches in the strike zone.
When thinking about plate coverage, consider that a baseball is about 2.9 inches in diameter. Joseph’s plate coverage of 81 square inches means he can hit and keep fair a baseball whose center of mass can be in about 28 different locations. Jose Abreu’s plate coverage of 66 inches means he can reach a baseball in about 23 different locations. Pitchers have more room to miss their spot with Abreu than they do with Joseph. These numbers aren’t literal; we’re looking at median absolute deviation and there’s a continuous range of baseball locations. But the general idea holds.
Here are the 2016 laggards:
This list contains up-the-middle players galore: Alex Avila, Josh Phegley, Greg Garcia, Francisco Cervelli, Tommy La Stella, Nick Hundley, and Delino DeShields. Motter is a utility player but has spent most of his time at shortstop. None of these guys are particularly good hitters. Now we may have an idea as to why. Pitchers have an easier time with these guys because they can’t reach as many balls. The guys in this list mostly rely on their defense to stick in the majors. Alex Avila’s coverage fits about 8.6 baseballs in it.
Comparing league-worst Avila to league-best Joseph, we see the differences in plate coverage:
Joseph clearly was able to hit, and keep fair, much wider range of pitches than Avila did.
Is More Plate Coverage Better?
But so what? Is increased plate coverage a good thing? The following graph gives us an idea:
The graph hints at an answer: “up to a point”. See how wRC+ rises along with plate coverage until the latter reaches about 50 square inches, then declines as coverage further increases. With an R2 of .019 vs. an R2 of .003, a parabola is a better fit than a linear model (although neither indicates plate coverage has a large effect on offense).
This answer makes sense. With too little coverage, pitchers can easily sneak pitches past you. With too much coverage, you put balls into play that you shouldn’t. There’s a sweet spot in the middle.
Recall the pitch DeShields hit, above. It was a clear ball, but he made contact with it anyway. Now see the following pitch to Caleb Joseph. It was 10.6 inches off the ground, the lowest pitch in 2016 that he hit fair, but he swung anyway:
That’s, er … not a pitch you should swing at. To be fair, there were runners on the corners with no one out in the seventh inning with the Orioles down by two. I’d be aggressive in that situation too!
Where’s the Sweet Spot?
The scatterplot above indicates the sweet spot is around 51 square inches. With this in mind, let’s see who had the most optimal plate coverage in 2016:
| Name | Distance from Optimal Coverage (sq. in) | wRC+ |
|---|---|---|
| Brett Gardner | 0.066 | 97 |
| David Peralta | 0.103 | 84 |
| Whit Merrifield | 0.113 | 89 |
| Zack Cozart | 0.114 | 91 |
| Evan Gattis | 0.116 | 119 |
| Ender Inciarte | 0.131 | 97 |
| Jose Altuve | 0.171 | 150 |
| Ryan Goins | 0.187 | 38 |
| Brandon Drury | 0.202 | 102 |
| Wil Myers | 0.212 | 115 |
These hitters averaged a 92 wRC+.
Now those with the least optimal plate coverage in 2016:
| Name | Distance from Optimal Coverage (sq. in) | wRC+ |
|---|---|---|
| Francisco Cervelli | 18.415 | 99 |
| Domingo Santana | 19.801 | 110 |
| Josh Phegley | 20.720 | 89 |
| Greg Garcia | 21.537 | 111 |
| Salvador Perez | 22.001 | 88 |
| Corey Dickerson | 24.931 | 101 |
| Keon Broxton | 25.548 | 109 |
| Taylor Motter | 25.609 | 66 |
| Alex Avila | 25.908 | 104 |
| Caleb Joseph | 30.354 | 6 |
Note this list is a mix of 2016 players with both large and small coverage areas. These hitters averaged an 88 wRC+, 4% worse than the group with optimal coverage.
Year-to-Year Changes
Now that we can estimate the optimal plate coverage for a hitter, we can see how year-to-year changes affect offensive performance:
On the x-axis, positive numbers indicate getting closer to optimal plate coverage. For example, a point at 10 could indicate the player moved from 30 square inches of plate coverage in 2016 (an optimal distance of 21 square inches) to 40 (optimal distance of 11 square inches) in 2017. Conversely, negative numbers indicate a player moved further away from optimal plate coverage.
The trend line indicates that the players who move closer to optimal plate coverage tend to see their offense increase. This trend supports the idea that more plate coverage isn’t always a good thing; rather, moving towards optimal coverage at around 51 square inches is. Of course, I’m not controlling for factors such as age, but the general idea seems to pass the sniff test.
Since last year, Jose Peraza’s expanded his plate coverage by over 30 square inches and in the process moved further away from the optimal area. His wRC+ has dropped 44 points.
Although it seems to be just one pitch, Peraza’s expanded his strike zone up.
Meanwhile Adam Jones has shrunk his plate coverage by 24 square inches. This shrinkage moves him closer to the optimal coverage area. His wRC+ is down only 3 points.
So far this year, Jones is either laying off high and down-and-in pitches or has been unable to put them in play. From what we’ve seen about plate coverage, this change may be beneficial.
Conclusion
Caveats abound here. Studying plate coverage doesn’t produce actionable advice. You can’t tell a player “You need to cover 10 more square inches of the strike zone.”
Further, we haven’t identified a causal relationship. Does a sub-optimal plate coverage cause a lower wRC+, or do players with less offensive prowess attempt to cover a sub-optimal area? Also, this study doesn’t control for age. A future study might get more granular by looking at xwOBA, instead of wRC+. Finally, I’ve used data to identify a major league-wide optimal plate coverage. Each player is bound to have his own optimal area, based on his skill set. This area may even vary by game situation.
But overall, studying plate coverage can give us a small bit of insight into why a player might be struggling at the plate. We now understand that more coverage isn’t always better. What matters more is how close the player is moving to the optimal coverage.
Ryan, I think this is extremely interesting. You’ve really tackled something that I think gets danced around a lot, but never truly tackled. Forgive me if this feels more critical than intended, but there are three things that I think you might want to consider (I’d love to see it revisited!):
1. I think your decision to omit foul balls wasn’t the best choice. In my opinion, the “purpose” of plate coverage is two-fold. First to lower the rate of called/swinging strikeouts, and second to get a good pitch to hit. By omitting foul balls you’re essentially penalizing players who battle around the edges until they get a better pitch to hit.
2. I think things might be improved by incorporating more context, especially pitch type and count. The CaJo example above is particularly heinous because it was on a 1-0 pitch, although it appeared to be a quality offering in his defense.
3. This may be covered by (2) but looking at swing discretion would be interesting. A player may appear to cover less of the plate when in fact they’re choosing to not swing at non-ideal pitches.
Again, I think this is terrific stuff. I would be very interested to see how/if things changed incorporating some of these pieces into a future analysis on the same topic!
Thanks Jeff! I don’t think you’re being too critical at all. You’ve made some really good suggestions and points.
This may be a philosophical difference we have about the definition of the word ‘coverage’. I understand your point but still don’t agree that fouling a ball off counts as ‘coverage’, even if it results in better pitches to hit because the plate appearance is extended. I’m just reporting balls hit fair; whether a batter can hit those pitches fair because they battled with a lot of foul balls is just outside the scope of this particular study.
I wonder if there’s a way to give like half-credit for foul balls. Or, perhaps a separate study of foul ball coverage is in order 🙂
Context is definitely an area I want to explore next. Eli suggested two-strike coverage; I think that’d be a good split to explore. I would imagine the context that most affects plate coverage is the current leverage index. I’ll note that for a future study.
Swing discretion as you describe it here seems translatable to ‘plate discipline’. This makes sense to look at in the future, for sure. It may be interesting to explore the difference between the two … who *thinks* they have great plate coverage but in reality should be more selective?
Thanks for the feedback and I’ll definitely let you know if I look at this again in the future 🙂
Do you have data on elite hitters like Trout, Miggy etc? Would be interesting to see what their figures are.
Glenn, Trout’s coverage varies between 53 – 57 inches (2015 – 2016). So, fairly close to ‘optimal’. Harper in his MVP season had a coverage of 51 inches. Cabrera covered 61 inches in 2015 and 49 inches in 2016.
Players with at least 160 wRC+ in a season had a mean distance from optimal coverage of ~4.2 square inches. Players with a wRC+ below 160 had a mean distance from optimal of ~7.2 inches. That’s a pretty coarse cut of data, but it’s more evidence supporting the optimal coverage at about 51 square inches.
Great stuff Ryan. Would love to see a follow up with 2-strike plate coverage. In my mind, before a batter has two strikes he should “zone-in” and try to hit his pitch, but with 2 strikes he would then need to “expand the zone” which is where plate coverage would be far more important.
Thank you! Two-strike coverage is a great idea for a follow-up; thanks.
I think part of your 0.02 correlation is a bias from where a player gets pitched. Bad hitter gets a lot down the middle, makes contact with them, those pull the coverage metric in. Good hitter gets pitches that avoid the middle, contact with those increases coverage.
Try smoothed contact rate as a function of horizontal position, and define a player’s coverage as the range where contact is over 50% or over total average or something.
Very good insight; I agree with the bias there. thanks!
Totally agree. You can’t tell a player “You need to cover 10 more square inches of the strike zone”
Ted williams said more talented hitters can cover more of the zone but a disciplined hitter that stays in his hot zone might be more productive than a high plate coverage hitter who expands.
Sometimes too much talent means you are too confident and swinging at everything.
But ideal would be a hitter who is patient and takes pitches at the periphery of the zone but still has plate coverage because that makes him less exploitable.
For example cabrera has a huge zone that he covers and he will expand some but generally doesnt go too far off the plate and forces the pitcher over the plate.