Prophetic Umpires

March 30, 2014
Posted by Jay Livingston

“It ain’t nothin’ till I call it,” said umpire Bill Klem. And if he called it a strike, a strike it was.  As Klem knew, the umpire has something resembling papal infallibility.  That was then. Klem worked behind the plate from 1905 to 1942 and holds the record for throwing players and managers out of the game (the infallibility thing is sometimes a bit much for players to take).  Now, thanks to modern technology, we can know just which calls the umpires miss.

Here’s Matt Holliday taking a called third strike.

Holliday’s body language speaks clearly, and his reaction is understandable. The pitch was wide, even wider than the first two pitches, both of which the umpire miscalled as strikes.* 

The PITCHf/x technology that makes this graphic possible, whatever its value or threat to umpires, has been a boon for sabremetricians  and social scientists.  The big data provided can tell us not just the number of bad calls but the factors that make a bad call more or less likely.  In the New York Times today (here), Brayden King and Jerry Kim report on their study of roughly 780,000 pitches in the 2008-09 season. Umpires erred on about 1 in every 7 pitches – 47,000 pitches over the plate that were called balls, and nearly 69,000 like those three to Matt Holliday.

Here are some of the other findings that King and Kim  report in today’s article.

•  Umpires gave a slight edge to the home team pitchers, calling 13.3% of their pitches outside the zone as strikes.  Visitors got 12.6%.
• The count mattered
•     At 0-0, the error rate was 14.7%.
•     At 3-0, 18.6% of pitches outside the zone were called as strikes
•     At 0-2, only 7.3% of pitches outside the zone were called as strikes
• All-star pitchers were more likely than others to get favorable calls . . .
• . . . Especially if the pitcher had a reputation as a location pitcher.
• The importance of the situation (tie game, bottom of the ninth) made no difference in bad calls.

It seems that expectation accounts for a lot of these findings. It’s not that what you see is what you get. It’s that what you expect is what you see. We expect good All-star pitchers to throw more accurately, especially control freaks like Greg Maddux.**  We also expect that a pitcher who is way ahead in the count will throw a waste pitch and that on the 3-0, he’ll put it over the plate.  My guess is that umpires share these expectations. The difference is that the umps can turn their expectations into self-fulfilling prophecies.

--------------------

* I took the graphics from fangraphs

**The pitcher in the clips is Tyler Clippard, a pretty good closer for the Nationals. He was selected as an All-star once, not nearly enough to meet the King-Kim criterion level of five.