Modern Baseball Statistics (Part Four – Ultimate Zone Rating)

A column by Eric Umphrey

This is part four in a series of articles introducing the new statistics that baseball announcers are using in broadcasts. Today's statistic is Ultimate Zone Rating which is used to measure a players defensive skill. UZR only started being recorded in 2002 so we don't have data before then. This statistic is provided by a company called Sports Info Solutions. They actually hire people to watch every defensive play in baseball and give it a rating. It looks to pay fairly well as the average salary for this position is $32,000 a year.

Before we had companies performing professional analysis all we had to go on for a player's defensive contribution was fielding percentage and errors. This left a lot to be desired since it told us nothing about a player's range in the infield or the outfield.

If a defender got a bad jump on the ball or was just slow, he rarely received an error on the play. Also, there was no account taken for how hard the ball was hit. A third baseman fielding a routine ground ball coming off the bat at 88 miles per hour was treated the same as a ball coming at 119 miles per hour.

Now that we have data being measured on each play and multiple people reviewing and scoring each play, we have a more informed idea on which players are good at defense. That said, it is still subjective and the numbers on individual players can vary significantly from year to year.

At a high level this is how plays are broken down into buckets and assigned values. Plays are judged to be bunts, ground balls, outfield line drives or outfield fly balls. The ball speed is then judged to be slow, medium or fast. The playing field is then broken up into zones and each zone is given a difficulty probability based on the type and the speed.

I'll walk through an example of how this works. Suppose a line drive is hit directly between the center fielder and right fielder. Each has an equal 50% chance at the ball but only a small, say 10%, chance to catch it. No one manages to catch it.

Let's say the average run value for the play is .82 runs. So .82 runs * 10% catch rate = .082 runs / each fielder is a -.041 runs taken off of their rating since they both had an equal chance. Now had it been caught let's say by the right fielder then the calculation would look like this. So .82 runs * 90% drop rate = .738 runs for the right fielder. The center fielder doesn't get penalized for not making the catch.

Here is a chart from Fangraphs with Mariners info added by me for context.

Rating UZR Player UZR Year

Gold Glove Caliber +15 Franklin Gutierrez 31.0 2009

Great +10 Brett Boone 10.0 2003

Above Average +5 Ichiro 5.0 2005

Average +0 Justin Smoak 0.7 2011

Below Average -5 Richie Sexson -4.8 2005

Poor -10 Brett Boone -9.7 2005

Awful -15 Raul Ibanez -20.3 2007

Here are the top six Mariners seasons for UZR.

Player UZR Year

Franklin Gutierrez 31.0 2009

Ichiro 21.1 2003

Ichiro 20.4 2004

Ichiro 20.3 2006

Mike Cameron 19.2 2003

Adrian Beltre 18.5 2006

Notice how Ichiro appears to have gone from gold glove caliber to just above average in 2005? Examples like this show how this statistic isn't perfected yet. There is work still to be done with this statistic, but it is much better than what we had before.