QuesTec and Umpires

One of the several benefits of SABR membership is the yearly publication of the The Baseball Research Journal. The journal contains a variety of aritcles on all aspects of baseball history and analysis written by SABR members. In this year's edition there were several interesting articles relating to the scientific aspects of the game and one in particular, "Cameras and Computers, or Umpires?" particularly caught my eye. The article was written by Robert K. Adair, whose The Physics of Baseball I've blogged about previously.

In the article Adair describes the QuesTec system that was installed in a handful of ballparks in 2002 to collect data on ball and strike calls in a system called the Umpire Information System (UIS). It was used in 10 ballparks in 2003 where the umpire is given a CD-ROM with all the pitches after each game. Ever since I heard how NFL officials are given a video tape after each game that scores them on each play I've thought that MLB should do something similar with umpires in order to better control umpires who have tendancies to large strike zones or to consistently give the tie to the fielder. Adair was able to analyze the stats from the system reported as follows:

# of pitches analyzed: 83,891
# of pitches where man and machine agreed: 71,164
# of pitches where they disagreed by < 2 inches: 4,970
# of pitches where they disagreed by > 2 inches: 7,757 (9% or 14 pitches per game)

Adair was then able to create a table that showed where the 9% of pitches disagreed:

Outside Corner  High

U-s Q-b 3,336   U-s,Q-b 301

U-b Q-s 122     U-b,Q-s 943

Inside Corner   Low

U-s Q-b 622     U-s,Q-b 18

U-b Q-s 208     U-b,Q-s 2,007

In the table U-s,Q-b would mean the umpire called it a strike and QuesTec a ball. Finally, from this data Adair was able to construct a graphic that shows the difference between the QuesTec strike zone and the umpires zone. The QuesTec zone is taller by 3/4 of a baseball at the top and over a full baseball at the bottom and thinner by a little less than that amount while the umpires zone is narrower and flatter and centered more towards the outside corner.

Certainly some of the differences may be attributed to errors in the system (reading this I was reminded of an article the journal of the IEEE Computer Society that described the K-Zone implementation for ESPN and noted how difficult it was to come up with an algorithm to accurately track the path of the pitched ball amidst the differences in lighting, weather, and fan apparel) but the model strike zones that Adair was able to construct seem to me to be exactly what a fan sees everyday. Pitches that are the width of a baseball outside are called strikes while pitches several inches below the letters are still balls. Part of this is certainly because an umpire typically leans over the catcher's left should and therefore does not have a good view of the outside corner and so more often has to guess at the actual location. This also explains why the smallest difference is on pitches on the inside corner where umpires have a great view already.

After seeing this data I'm more convinced than ever that this sort of system can and does work and should serve to help umpires call the stirke zone found in the rule book.

"The STRIKE ZONE 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."

By doing so it will minimize one of the three strike zones described by Ted Williams:

"The batter [and subsequently the pitcher] has three strike zones: his own, the opposing pitcher's, and the umpire's. The umpire's zone is defined by the rule book, but it's also more importantly defined by the way the umpire works. A good umpire is consistent so you can learn his strike zone. The batter has a strike zone in which he considers the pitch the right one to hit. The pitchers have zones where they are most effective. Once you know the pitcher and his zone you can get set for a particular pitch."