Ballpark Strike-Zone Factors

So, look, I feel like I need to explain why I’m looking at something you’ve probably never even thought about before. Obviously, I write a lot about pitch-framing, and about the different strike zones that different players get. Some time back, I was writing about Zack Greinke‘s favorable zone with Milwaukee. As such, I wrote about Jonathan Lucroy‘s receiving, and I wrote about Martin Maldonado‘s receiving. Afterward, I received an interesting idea from a player:

I’m convinced the background in Milwaukee affects the home plate umpires’ depth perception and expands the zone down and away. Is there a home road split for both of the catchers? Maybe that will make some sense for my possibly distorted perception of balls and strikes in Milwaukee.

Not something I’d considered. Made sense, in theory. Perhaps a hidden park factor was inflating the framing numbers. Maybe Milwaukee is like the Colorado of expanding the strike zone. For whatever reason I never got around to researching this, until this afternoon, when I compiled the relevant data.

Per usual, I’m thankful for the existence of Baseball Savant. I looked at called-strike rates both inside and outside of the zone for 2012 and 2013, for every team, for hitters and pitchers, home and away. I combined the data that needed to be combined and calculated ratios of home strike rates to road strike rates. It’ll all make more sense with the data and a bit of explanation, so let’s just get to that part now. Here’s a table of home-to-road called-strike-rate ratios:

Team In-Zone Out-Of-Zone
Angels 1.03 0.98
Astros 1.01 0.96
Athletics 1.00 0.99
Blue Jays 0.98 0.98
Braves 1.00 1.02
Brewers 1.01 0.99
Cardinals 0.99 0.97
Cubs 1.00 1.03
Diamondbacks 1.00 0.97
Dodgers 0.99 0.97
Giants 1.00 1.00
Indians 1.02 1.02
Mariners 1.01 1.06
Marlins 0.99 1.01
Mets 1.00 0.96
Nationals 1.00 1.01
Orioles 0.99 0.93
Padres 1.02 1.06
Phillies 1.01 1.00
Pirates 1.00 1.01
Rangers 0.97 1.06
Rays 1.02 1.01
Red Sox 1.00 0.97
Reds 1.02 1.01
Rockies 0.98 0.96
Royals 1.00 0.98
Tigers 1.01 1.04
Twins 1.01 0.99
White Sox 0.97 1.03
Yankees 1.00 1.07

It ought to be sortable! What do we find? Let’s look at the Brewers and Miller Park first, since they inspired the whole idea. The last two years, in Miller Park, called strikes have been up one percentage point within the zone. Outside of the zone, though, they’ve been down a tiny bit. The difference is small, but overall I’m left unconvinced by the player’s theory — it doesn’t look like Miller has a huge park factor at play. Perhaps there’s slight strike inflation, but it doesn’t explain Lucroy and Maldonado away.

Other numbers are more interesting. In Angels home games, 90.0% of pitches taken in the zone have been called strikes. In Angels road games, 87.8% of pitches in the zone have been called strikes. Yet when it’s come to pitches out of the zone, umps have been a little more stingy in Anaheim. I’m not sitting here, offering explanations. I’m just offering numbers and you can try to come up with your own explanations.

At the other extreme, Rangers and White Sox home games have featured reductions in in-zone called strikes. Yet they’ve also featured slight increases in out-of-zone called strikes. What about ballparks where both factors are either above or below 1.00? St. Louis, Baltimore, Los Angeles, Toronto, and Colorado have reduced both kinds of called strikes. Seattle, San Diego, Cincinnati, Cleveland, Tampa Bay, and Detroit have increased both kinds of called strikes.

In Yankee Stadium, 16% of pitches taken out of the zone have been called strikes. In Yankees road games, that drops to 15%. In Camden Yards, just under 14% of pitches taken out of the zone have been called strikes. In Orioles road games, that jumps to nearly 15%. Do teams bat more lefties in New York? Does that open them up to the lefty strikes away off the plate? I haven’t a clue. Again, I’m just playing the part of messenger.

In Coors Field, in-zone strikes have been down 2.1 percentage points, and out-of-zone strikes have been down 0.5 percentage points, for a sum effect of -2.6 percentage points. That ties the park with US Cellular. Flipping around, we have Petco Park, where in-zone strikes have been up 1.8 percentage points, and out-of-zone strikes have been up 0.9 percentage points, for a sum effect of +2.7 percentage points. Next-closest is Angel Stadium, at +1.9. The Padres have come away recently with some surprisingly terrific pitch-framing numbers. Might some of that have been the ballpark, somehow? I don’t know nearly enough to be able to answer that, but I’m at least open to the concept. There can be park factors for anything.

The player at the start had an idea about Miller Park. After digging in, I don’t see a strong effect, but I’m glad to have had the idea provided to me, and there seem to be stronger effects elsewhere. Is there something about San Diego that makes an umpire more likely to call a strike? Is there something about Colorado or Chicago that does the opposite? These are things to investigate, as never does the tunnel of questions end. It gets narrower in places, sure, but we’ll always be able to keep going.



Print This Post



Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.


Sort by:   newest | oldest | most voted
sambf
Member
sambf
2 years 4 months ago

What’s the relative size of these numbers and pitch framing numbers? If they’re all signal and no noise, how much of pitch framing could they explain?

On the other hand, what’s the SD of those numbers, and what would you expect just from noise?

novaether
Member
novaether
2 years 4 months ago

I second this question. It looks like nearly all noise to me, given that:

1) If you were to bucket the parks into above/below average in/out of the strike zone (in a 2×2 grid) it would be pretty evenly distributed. This would contradict any overall strike-factor hypothesis.

2) The ratios for in-zone strikes are more closely bundled than out-zone strikes. This makes sense if these numbers were randomly generated because the denominator is smaller in the out-zone strikes calculation.

Benjamin
Guest
Benjamin
2 years 4 months ago

I feel like you are a step or two ahead of yourself. Has anyone really dug into park factors for Ks and BBs?

Jonah Pemstein
Member
Member
2 years 4 months ago

Yes. You can find them here:

http://www.fangraphs.com/guts.aspx?type=pf&teamid=0&season=2012

There isn’t too much variance in those numbers either, but there is some.

Benjamin
Guest
Benjamin
2 years 4 months ago

I know that they’re available. I want to know if anyone has explored them and figured out why they exist.

Joe
Guest
Joe
2 years 4 months ago

What if the discrepancies are caused not by the umpires, but by the PITCHf/x systems in each park? If the whole strike zone, or even just some part of it, were off by a few millimeters or whatever amount, it seems like that could have a similar effect.

JKB
Guest
JKB
2 years 4 months ago

It depends on whether the bias in a park’s PITCHf/x system is constant or varies randomly around the “true” strike zone from game to game. At least it is the same PITCHf/x system for every game in a ballpark. Umpires are more complicated. Unless MLB assigns each umpire to the same number of games behind the plate in each stadium, a compounded bias can arise if the umpires that call the games in one stadium share the same type of bias relative to the PITCHf/x system there.

Alaric
Guest
Alaric
2 years 4 months ago

I’m thinking that would lead to results like those observed for the Rangers and White Sox?

A lower percentage of “in zone” strikes (because some of them aren’t actually in the “real” zone) and a higher percentage of “out of zone strikes” (because some of them are actually in the “real” zone).

Matt
Guest
Matt
2 years 4 months ago

Did you seperate by zone? It’d be interesting to try to verify the low-and-away claim that the player made about Miller Park. Also, do 2012 numbers resemble 2013 at all?

MGL
Guest
MGL
2 years 4 months ago

I too, thought about how this could be distinguished from pitch f/x calibration errors.

And you definitely want to at least control for pitchers, batters, and especially umpires. If you don’t, that contributes to the non-park related variance (noise).

For the noise (non-park related variance), you have random variance which is a function of sample size, plus the umpire variance, plus the batter variance, plus the pitcher variance, plus the catcher framing variance (the same exact catchers are not catching at home and away, for both teams), plus whatever else might cause differences in strike zone accuracy, like weather, day/night games, etc.

That’s a lot of noise!

MGL
Guest
MGL
2 years 4 months ago

The thing is, a good catcher framing model would control for pitchers, batters, and umpires, at the least, and if park effects were due to those same things, then you would not want or need to adjust the catcher framing numbers for park effects, as that would be “double adjusting.”

For example, let’s say that in MIL, you happened to have had umpires with large strike zones as compared to the other parks. So, everything else being equal, it would look like MIL was a “large strike zone” park and you might be tempted to adjust Lucroy’s framing value downward. But, if his numbers are already umpire adjusted, his raw numbers would already be adjusted.

AK7007
Member
AK7007
2 years 4 months ago

Hm, I wonder if there is any kind of correlation between parks that have, for example grassy areas around the hitter’s backdrop instead of bleachers. (SD has that picnic hill) Or crazy video boards. Or covered vs entirely outdoor stadiums. Day/Night splits? There are so many qualitative differences (also known as” categories!”) that could be used to sort parks into groups and see if there are things that muck with strike calling. And I think when it’s not late at night I might mess around with, unless Jeff takes it upon himself to do the work that likely won’t pan out for me.

pft
Guest
pft
2 years 4 months ago

Could the variance be at least in part due to the umpires rotations. I have to imagine the umpires are not evenly distributed among all 30 parks, and some crews work certain parks more often.

The only numbers that striked me as being significant where the Mariners and Yankees out of zone calls. I wonder if King Felix and CC/Rivera were responsible for some of that, having a reputation may have helped them get calls.

Nathaniel Dawson
Guest
Nathaniel Dawson
2 years 4 months ago

“At the other extreme, Rangers and White Sox home games have featured reductions in in-zone called strikes. Yet they’ve also featured slight increases in out-of-zone called strikes.”

Hmmm. Sounds like the Umps somehow don’t see pitches quite as well in those ballparks? Like you say, there’s so much about park factors we don’t know yet. I’m not convinced that a lot of this isn’t just noise, but you’ve certainly stumbled onto something that’s very intriguing.

“In Coors Field, in-zone strikes have been down 2.1 percentage points, and out-of-zone strikes have been down 0.5 percentage points, for a sum effect of -2.6 percentage points.”

As if pitchers don’t have it tough enough there already.

failed mathematician
Member
failed mathematician
2 years 4 months ago

White Sox announcers regularly allude to the difficulty seeing the ball as a batter during evening games as a shadow passes over the mound. So the pitch is thrown under sunlight and passes into the shadow. Granted one of those announcers is hawk “twtw” harrolson. But one of the radio guys Darrin Jackson says he experienced it in his time as a player. Sounds like there’s something to it and it is effecting umpires too.

tangent
Guest
tangent
2 years 4 months ago

Jeff, did you correct for count here?

We know the size of the zone varies a lot in a hitter’s count versus a pitcher’s count. We know that parks have BB and K factors, so I figure they probably have factors for how many pitches are thrown in which count. This could be some of what you’re seeing (beyond noise).

racehorse1
Guest
racehorse1
2 years 4 months ago

Why reduced strikes in st. Louis? Watch the games this year, and see how the umpires refuse to call a 3rd strike against the cardinals. And watch how the cardinals go insane, the few times they do get rung up. Its blatant.

Poach
Guest
Poach
2 years 4 months ago

One approach would be to look at whether the park factor for a given ballpark is consistent from year to year, or whether it varies wildly, indicating that this is mostly noise (which is my guess).

evansimon
Member
evansimon
2 years 4 months ago

As far as Miller Park goes, they led the league in walks last year, and have been in the top 10 in walks from 2010 to 2013. This according to espn’s park factor.

http://espn.go.com/mlb/stats/parkfactor/_/year/2013/sort/walksFactor

wpDiscuz