2011 Umpire Projections

An umpire is not supposed to have any influence on any game, but many times they do, especially the home plate umpire calling balls and strikes. Even though the strike zone is supposed to uniform across the league, each umpire has their own unique strike zone. I have gone back over the past 3 years and projected how pitcher or hitter friendly each umpire will be for the up coming season.

First, I went back and took the 2007 to 2009 data and looked to see how much to weight each year’s data to project the known 2010 data. I proved common sense and an even 1/3 weighting of each year as the best predictor. Along with the weighting, I also regressed each umpire to the league average. Finally, I took the 2008 to 2010 values and created a projection for how far the umpire’s zone differs from the league average. The complete list is at the end of the article.

Besides the actual values, here is a look at how pitcher-friendly umpire Brian Runge’s and hitter-friendly umpire Jerry Crawford’s strike zones compare using heat maps for both right- and left-handed pitchers.

The image is from the catcher’s and umpire’s perspective, the box being the rule book strike zone and the circle is used just for reference. The scale on the right is the percent, in decimal format, of pitches called strikes. The data is for all pitches from 2008 to 2010.

LHP
Brian Runge


Jerry Crawford

RHP
Brian Runge

Jerry Crawford

Their differences aren’t huge, but if you look at the green zone, the 50-50 line, Runge expands the zone out a bit. So, if during this season you feel an umpire is squeezing your favorite pitcher a bit, look to see if it’s the umpire’s normal tendency to have a small zone. If not, he may just hate your team.

———————————————————————————————-

List of umpires by how much they are pitcher (positive % Difference compared to Mean) or hitter (negative % Difference compared to Mean) friendly.

Columns are as follows:
Name – The name their mother calls them, not “You Dirty Bum” – we try to keep it P.G. around here.
% of Called Strikes in Strike Zone – Percentage of pitches in the rule book strike zone called strikes.
% Difference compared to Mean – % of how much larger the individual umpire’s strike zone is compared to the league average.
Reliability – How much of the umpire’s projection is from their own values. The rest is how much regression is added in. For example David Rackley’s has a reliability of 0.70, so 70% of the projection is based on pitches he called and 30% is the regression amount.

Name % of Called Strikes in Strike Zone % Difference compared to Mean Reliability
Rackley, David 25.9% -6.6% 0.70
Larson, Barry 25.9% -6.5% 0.26
Tumpane, John 27.7% -4.0% 0.66
Crawford, Jerry 27.8% -3.8% 0.96
Bakke, Chris 27.9% -3.7% 0.30
Merry, Jon 28.4% -3.0% 0.54
Loveless, Eric 28.5% -2.8% 0.27
Tiller, Chris 28.7% -2.5% 0.82
Schrieber, Paul 28.8% -2.5% 0.97
Barrett, Lance 28.8% -2.4% 0.31
Hohn, Bill 28.8% -2.4% 0.97
Conroy, Chris 29.1% -1.9% 0.40
Johnson, Adrian 29.2% -1.9% 0.98
Hickox, Ed 29.2% -1.8% 0.96
Campos, Angel 29.6% -1.3% 0.97
Cousins, Derryl 29.6% -1.2% 0.98
Rapuano, Ed 29.6% -1.2% 0.98
McClelland, Tim 29.7% -1.1% 0.98
Davis, Gerry 29.7% -1.1% 0.97
Tschida, Tim 29.8% -1.0% 0.97
Layne, Jerry 29.8% -1.0% 0.97
Danley, Kerwin 29.8% -1.0% 0.96
Marquez, Alfonso 29.8% -1.0% 0.96
Gibson, Greg 29.9% -0.9% 0.97
Barry, Scott 29.9% -0.8% 0.97
Tichenor, Todd 30.0% -0.7% 0.97
Hoye, James 30.0% -0.6% 0.98
Knight, Brian 30.0% -0.6% 0.97
Fletcher, Andy 30.1% -0.6% 0.97
Carapazza, Vic 30.1% -0.5% 0.75
Reyburn, D.J. 30.2% -0.4% 0.93
Fairchild, Chad 30.2% -0.4% 0.97
Wegner, Mark 30.2% -0.3% 0.97
Welke, Tim 30.2% -0.3% 0.97
Iassogna, Dan 30.3% -0.3% 0.97
Uyl, David 30.3% -0.3% 0.96
Muchlinski, Mike 30.3% -0.3% 0.93
Beal, Damien 30.3% -0.3% 0.85
Wendelstedt, Hunter 30.3% -0.2% 0.97
Joyce, Jim 30.3% -0.2% 0.97
Reilly, Mike 30.3% -0.2% 0.96
Dreckman, Bruce 30.4% -0.2% 0.97
Scott, Dale 30.4% -0.2% 0.97
Guccione, Chris 30.4% -0.1% 0.97
Hallion, Tom 30.4% -0.1% 0.97
Bell, Wally 30.4% -0.1% 0.98
Welke, Bill 30.4% -0.1% 0.97
Fagan, Clint 30.4% -0.1% 0.25
Estabrook, Mike 30.4% 0.0% 0.97
Holbrook, Sam 30.5% 0.1% 0.97
Carlson, Mark 30.6% 0.1% 0.96
Winters, Mike 30.6% 0.2% 0.97
Meals, Jerry 30.6% 0.2% 0.98
Foster, Marty 30.7% 0.3% 0.97
DeMuth, Dana 30.7% 0.3% 0.97
Blaser, Cory 30.7% 0.4% 0.51
Everitt, Mike 30.7% 0.4% 0.98
Hudson, Marvin 30.7% 0.4% 0.97
Kellogg, Jeff 30.7% 0.4% 0.97
Cederstrom, Gary 30.7% 0.4% 0.97
Culbreth, Fieldin 30.8% 0.4% 0.98
Kulpa, Ron 30.8% 0.4% 0.97
Robinson, Will 30.8% 0.5% 0.96
Bucknor, CB 30.8% 0.5% 0.98
Barksdale, Lance 30.8% 0.5% 0.97
Porter, Alan 30.8% 0.5% 0.79
Darling, Gary 30.8% 0.5% 0.97
Barrett, Ted 30.9% 0.6% 0.97
Timmons, Tim 30.9% 0.6% 0.97
West, Joe 30.9% 0.6% 0.98
Wolf, Jim 30.9% 0.6% 0.97
Reynolds, Jim 30.9% 0.7% 0.97
Hernandez, Angel 30.9% 0.7% 0.97
Vanover, Larry 31.0% 0.7% 0.92
Nelson, Jeff 31.0% 0.7% 0.97
Gonzalez, Manny 31.0% 0.8% 0.96
Drake, Rob 31.0% 0.8% 0.98
Davidson, Bob 31.1% 0.8% 0.97
Nauert, Paul 31.1% 0.9% 0.97
DiMuro, Mike 31.3% 1.2% 0.97
Diaz, Laz 31.3% 1.2% 0.97
Cuzzi, Phil 31.4% 1.4% 0.97
Bellino, Dan 31.5% 1.4% 0.93
O’Nora, Brian 31.5% 1.4% 0.97
Hirschbeck, John 31.6% 1.7% 0.95
Eddings, Doug 31.7% 1.8% 0.97
Emmel, Paul 31.8% 1.9% 0.97
Randazzo, Tony 32.1% 2.3% 0.96
Gorman, Brian 32.1% 2.3% 0.93
Cooper, Eric 32.1% 2.4% 0.97
Ripperger, Mark 32.2% 2.4% 0.53
Miller, Bill 32.3% 2.6% 0.98
Runge, Brian 32.9% 3.5% 0.89


Print This Post



Jeff writes for FanGraphs, The Hardball Times and Royals Review, as well as his own website, Baseball Heat Maps with his brother Darrell. In tandem with Bill Petti, he won the 2013 SABR Analytics Research Award for Contemporary Analysis. Follow him on Twitter @jeffwzimmerman.


Sort by:   newest | oldest | most voted
ryanmira
Member
ryanmira
5 years 2 months ago

Do you have projections on umpire ejections as well? I’ve been in shambles since Bruce Froemming retired.

rbt
Guest
rbt
5 years 2 months ago

Yes, we need that for the Umpire Ejection Fantasy League.

I think we also need to have a projection of how many balks Joe West will call on Mark Buerhle.

UZR is a Joke
Guest
UZR is a Joke
5 years 2 months ago

Since Buerhle balks on every pickoff, hopefully a lot.

bill
Guest
bill
5 years 2 months ago

Am I reading this right? Most umpires only call about 30% of “rule book” strikes actual strikes?

Rob
Guest
Rob
5 years 2 months ago

I was wodering the same thing. That is a small number.

MikeM
Guest
MikeM
5 years 2 months ago

Yeah, they really call the rulebook zone about as well as NBA refs adhere to by-the-book travelling violations.

Do remember though, that hitters are more likely to swing at pitches right down the pipe, so the “called strike” bucket is going to contain more pitches in the outskirts of the zone than pitches in the dark red center.

Articles like these just harden my position that we need to make use of the technology availabe to us and use the camera system to call balls and strikes. Free the homeplate umpire up to focus on check swings and enforcing a strict time limit on batters stepping out of the box.

Jcantwell
Guest
Jcantwell
5 years 2 months ago

No, that can’t be. Using the green circle for the 50/50 mark, that covers most of the strike zone, so anything inside that hey are calling even higher. I think that column means something different?

Telo
Guest
Telo
5 years 2 months ago

Yea, it’s wrong. No idea what exactly what that number is supposed to represent. He may have just calculated the rulebook zone to be massive, but there’s just no way that 30% is remotely correct as the average.

Millsy
Member
Millsy
5 years 2 months ago

No, I think what it says is this:

Of the pitches called strikes, 30% of them are within the rulebook zone.

This means that there are lots of pitches on the edge of the strike zone (just outside it) thrown by the pitchers. The majority of called strikes are at the edge of the zone because the batter usually waits to swing at one more in the zone. Any pitch swung at by the batter is not a called strike, therefore the numbers are lower than you would expect.

I think that perhaps the tally should go a different way: percent of pitches WITHIN the zone that are correctly called a strike. I’ve been working on something very similar and the range for this variable is something like 82% to 92%.

So, given a pitch within the strike zone that the batter does not swing at, the umpire is going to call it a strike (not worrying about count and pitch types) about 87.5% of the time on average (that’s 2010 data).

Telo
Guest
Telo
5 years 2 months ago

That would be incredibly difficult to believe. The rule book strikezone is:

“that area over home plate the upper limit of which is a horizontal line at the midpoint between the top of the shoulders and the top of the uniform pants, and the lower level is a line at the hollow beneath the knee cap. The Strike Zone shall be determined from the batter’s stance as the batter is prepared to swing at a pitched ball.”

If he correctly applied the rulebook zone to his data, there is no way “Of the pitches called strikes, 30% of them are within the rulebook zone.” With no data to back me up, I am calling that literally impossible with 100% confidence.

Millsy
Member
Millsy
5 years 2 months ago

Telo,

You might be right. My numbers for 2010 tell me that–of called strikes–65% of them were within the zone on average.

Telo
Guest
Telo
5 years 2 months ago

That definitely sounds reasonable. Jeff made an error somewhere – maybe applying the rulebook zone, computational, or simply mislabeling his table – but there is something screwy going on.

Millsy
Member
Millsy
5 years 2 months ago

Here’s what I came up with:

http://princeofslides.blogspot.com/2011/03/umpire-strike-zones.html

I apologize for the shameless (shameful) self-promotion.

Telo
Guest
Telo
5 years 2 months ago

So, he probably didn;t make a mistake. Just a really poor and lazy way to present the data.

Albert
Guest
Albert
5 years 2 months ago

Maybe some are swinging strikes, and therefore not “called” strikes? I agree, the way it is written doesn’t make much sense.

Telo
Guest
Telo
5 years 2 months ago

Put a buzzer in umpires pockets. It buzzes for a strike, no buzz for a ball. Don’t tell hitters.

Baseball just got better.

Derek R-C
Guest
Derek R-C
5 years 2 months ago

I am not sure if technology is that quick yet. I think it would take a second or 2 for the computers to calculate whether it was a strike or not and then communicate that to the umpire’s device.

Then the umpire actually gets to react to it. Hitters would wonder why it is taking the ump 2 seconds longer on every pitch to call ball or strike.

Millsy
Member
Millsy
5 years 2 months ago

Good grief, Jeff. You beat me by a day on this one! Good stuff. I guess my next post will be old news…

Matt Defalco
Guest
Matt Defalco
5 years 2 months ago

I will not lie. As a notgraphs reader, I went straight for Joe West.

steex
Member
steex
5 years 2 months ago

As did I. Had to make sure he had an acceptable rating, otherwise Joe might be forced to eject it.

Navin Vaswani
Member
5 years 2 months ago

You both are the best.

Peter Jensen
Guest
Peter Jensen
5 years 2 months ago

Jeff – I thought the rule book strike zone is plus or minus 10 inches in the X dimension, not 1 foot as it looks like you have drawn it.

eastsider
Guest
eastsider
5 years 2 months ago

Jeff,

I really like your umpire heat map series (I remember one back around the world series). I assume that the box you have drawn as the strike zone is largely for illustrative purposes and that you use the sz_top and sz_bot data in pitchfx for actual strike zone purposes…

Also, going to Peter Jensen’s question above, what do you use for the width of the strike zone. Home plate is 17″ wide – 8.5 in each direction from the middle. The ball is a little under three inches wide. So assuming a little bit of the ball catches a little bit of the plate, it seems like the strike zone should be about a little less than 11.5 inches wide.

Thanks

Telo
Guest
Telo
5 years 2 months ago

11.5 x 2 = 23 inches = the generally accepted notion of the strike zone.

eastsider
Guest
eastsider
5 years 2 months ago

Whoops. Thanks for doubling that value back for me. I should have said 11.5 inches wide in each direction.

Al Dimond
Guest
5 years 2 months ago

I’m not coming at this with any insider knowledge, so I might be missing something or ignoring some prior authoritative statement… the horizontal location of the pitch must refer to the middle of the baseball. So the location can only be 1.5″ off the corner for part of the ball to cross part of the plate. Wouldn’t the width of the strike zone therefore be 20″?

eastsider
Guest
eastsider
5 years 2 months ago

I guess that is what I am driving at. I don’t know if it is any part of the ball that has to cross any part of the plate or the middle of the ball, the entire ball, etc. When I pitched as a kid, I didn’t think about these things. I just assumed if I threw it, they should call it a strike, darn it!

JR
Guest
JR
5 years 2 months ago

My interpretation is that of all pitches in the strike zone,including swinging strikes,these are the percent of those that are called strikes. It would probably be better to just eliminate the swinging strikes from the mix,but the rankings would be essentially the same…

Telo
Guest
Telo
5 years 2 months ago

Well, I think you’re on the right track… but I think it must be ALL pitches… which in the end, should’ve been obvious I suppose, if not extremely misleading and borderline stupid to post the data that way.

% of Called Strikes in Strike Zone – Percentage of pitches in the rule book strike zone called strikes.

Really means, % of all pitches, including balls in play, swings and misses, etc, that result in a called strike by the umpire.

Telo
Guest
Telo
5 years 2 months ago

Sorry, my definition is missing a phrase:

% of ALL pitches thrown within rulebook strikezone, including balls in play and whiffs, that result in a called strike by the umpire.

Fergie348
Guest
Fergie348
5 years 2 months ago

If the rulebook strike zone were to be called consistently in MLB, it would revolutionize the game. I don’t think anyone is ready for such an event, and judging by the behavior of the main stakeholders it will never happen, at least not in my lifetime.

Nate
Guest
Nate
5 years 2 months ago

Fascinating stuff.

I’m a bit confused about the % called strike stat as well. Judging by the low percentage results, my first thought was that it was only calculating pitches that WERE NOT swung at, fouled off, or put in play, but merely the pitches that were looked at by the batter and thus left to the behest of the umpire to judge.

Is that correct?

Also, what’s the SD for distance from the mean? Any sort of normalization going on?

Kazinski
Guest
Kazinski
5 years 2 months ago

Millsey’s data at Prince of Slides has this amazing tidbit, if true. It’s not that I doubt his work but it is stunning.

Left handed batters get an enormous edge on incorrect strike calls over right handed batters. League wide left handed batters get 17.6% of pitches in the zone called as balls, for right handers get 14.2% of called pitches in the zone are called balls. That is a 3.4% differential, and it seems to be a consistent bias accross

I don’t know what the league wide walk percentage is for left handers is as opposed to right handers, but I would imagine left handers walk more and hit for a higher average, as a result of umpire bias for LHB.

The absolute worse is Chris Conroy with a13.85% incorrect RHB strike% vs. 37.04% incorrect LHB strike%.

Millsy
Guest
5 years 2 months ago

Kazinski,

I suspect it is the case, but there is a caveat. I believe is well known that the LHB zone extends further outside the ‘rulebook’ zone. However, it is called consistently, so a lefty may crowd the plate to take advantage (or it is called that way BECAUSE lefties may crowd the plate more…Mike Fast has talked about stances and how it affects the strike zone in the past).

This is why the strike zone is generally considered 2 feet across, rather than the book zone: it just extends out beyond there. So the numbers should be taken with a grain of salt.

As for Conroy, he is one of the umps with a very small sample size. I’ll try to include the sample sizes tonight as well. Sorry for that confusion.

Millsy
Guest
5 years 2 months ago

Forgot to say that the above reasoning means that the balls inside on lefties are not called strikes as much. The whole zone–on average–is just shifted away.

MGL
Guest
MGL
5 years 2 months ago

Are the numbers in column 3 the projections or the actual? IOW, do they include the regression or no?

Ian
Guest
Ian
5 years 2 months ago

Thanks for providing this information. As a fan, it really bothers me that there isn’t a standard. I can only hope that the umpires or MLB look at this stuff and realize that they need to change. Seeing bad calls makes many fans angry, please make it stop or at least happen less. This is a situation that can be improved.

adohaj
Guest
adohaj
5 years 2 months ago

Wow left handed batters better start swinging at pitches that are outside

fothead
Member
fothead
5 years 2 months ago

Wouldn’t it be hard to determine this on an across the board basis when every player’s strike zone is a different height?

Also about “the high strike” not being called.

“…The Strike Zone shall be determined from the batter’s stance as the batter is prepared to swing at a pitched ball.”

I’ve heard this explained by umpires to mean when a hitter is at the earliest part of a swing, not when they are just standing there waiting for a pitch. Regardless of stance, most hitters get to a very similar “ready” position which is what I’ve always understood the strike zone to be measured against. The rule book seems to say that but is somewhat unclear. But if you look at it that way, hitters will usually bend at the knees and waist a bit making the strike zone at or barely above the belt when they are standing upright.

If this were not the case, then a player could come to the plate with their chest against their knees (which would look rediculous) never swing and would probably walk 300 times in a season.

David K
Guest
David K
5 years 1 month ago

That’s amazing to me that the umps get the calls “right” only 1/4 to 1/3 of the time.

What I would have liked to see in the table is not just the % of pitches that are called strikes that are actually in the strike zone, but also the inverse — the % of pitches that are called BALLS that are actually in the strike zone. And then add another column for average runs scored per game to see what the correllation is between the umps strike zone and scoring. The umps would have to have a reasonable distribution of pitching talent and hitter vs. pitcher friendly ballparks to draw fair conclusions from the data though.

wpDiscuz