Zone rating has been around for 20 years now as a way to measure defensive players. It’s a simple concept; you just need to know two things: How many plays a fielder makes, and how many were hit into his zone. You divide the first number by the second, and you have zone rating. You can look at how many plays were made compared to an average fielder to get a plus minus rating, and that can easily be converted to runs to help us measure the overall worth of a ballplayer.
There’s More than one Zone Rating?
Last winter John Dewan and Baseball Info Solutions published The Fielding Bible and introduced a new version of zone rating. Dewan happens to be the guy who came up with the old Zone Rating when he worked for STATS, and he was able to add an improvement in the treatment of plays fielded outside of the zone.
In the original Zone Rating, if a ball is hit outside of your zone but you range far enough to make a play anyway, that ball is added to both your plays made and your chances. The effect of this is to underrate players with outstanding range. Here’s an illustration:
Take two shortstops, call them Billy and Jason. Both players have three ground balls hit near them, two in the shortstop zone and one just beyond it, hit straight up the middle. Jason makes plays on the two hit in his zone but cannot reach the other ball. Billy makes one play in his zone, makes an error on the other, but makes a great play on the ball hit up the middle.
The zone rating for Jason is a perfect 2/2. For Billy, its 2/3, even though both players have had the same opportunities, and recorded the same number of outs. The new Zone rating treats these plays differently. Balls in zone counts only those hit into your zone, and there is a separate category for balls fielded outside your zone.
Let’s Check the 2006 Results:
The new Zone Rating is now available on The Hardball Times website. THT has data for the 2004 to 2006 seasons, and will add 2007 data as soon as the games begin. To get a feel for how the new feature, balls out of zone, affects a player’s rating, let’s take a look at the shortstops for 2006, using the new data but calculated the same way as the old Zone Rating:
BIZ: Balls in zone
Plays: Plays made in zone
ZR: Zone rating
OOZ: Plays made out of zone
Chances: BIZ + OOZ
Total Plays: Plays + OOZ
Plays above average is found by finding the league average ZR, multiplying by Chances, and comparing to total plays made. This is for illustration purposes; I am using the total of this group of qualified players instead of the actual league average which would include starters and reserves. In addition, Chris Dial uses (correctly in my opinion) the AL and NL averages instead of an overall MLB average. There are factors (such as pitcher hitting) that make within league comparisons more accurate than overall comparisons.
Now, let’s try this again, treating balls outside of zone separately, so that we correctly reward both reliable fielders in their zone and players who make more plays outside of their zone. I will find the league average using only BIZ and Plays, and then do a separate calculation for balls outside of zone:
We can compare the player’s OOZ plays to the average OOZ plays based on how many balls were hit in their zone. This is similar to what David Gassko and Chris Constancio did for defensive ratings in the Hardball Times Season Preview. For this to work we have to assume that the number of out-of-zone opportunities for players are proportional to their in zone opportunities. This may or may not be a good assumption. An alternative would be to use total ground balls as the denominator. A better way would be to create a measure of out of zone opportunities. For a shortstop we could take every groundball hit to the left of the third baseman or to the right of the second baseman, and subtract the ones already counted in the SS zone. This measure is not available, so for now I think using BIZ as an estimate is acceptable.
“Reliability” is a measure using only balls in zone. For this group of shortstops, the average Zone Rating was .825. This group also made .1274 plays out of zone for every ball hit into the zone. “Range” compares the player’s out of zone plays made to this average, and the new total is simply Range + Reliability.
Jason Bartlett did very well in Zone Rating but made very few plays outside his zone. In the end, he’s just an average shortstop.
Bill Hall at the first step look a little below average, but made a tremendous number of plays outside his zone. This year he is being moved from shortstop to center field, but this may be a mistake as he makes many more plays than the average shortstop.
Derek Jeter looks a little below average at first, but he’s even worse than that. Jeter has very limited range. Only Bartlett made fewer plays outside the zone, but Jeter played almost 50% more innings than Bartlett.
Hanley Ramirez looks below average at first, tied with Jeter, but made a great number of plays out of his zone. He’s actually average or slightly above.
Adam Everett already ranked far above any other shortstop in the game, but in addition to fielding everything hit near him he has excellent range as well. He may be the best defensive shortstop since Ozzie Smith.
New Zone Rating vs. Old Zone rating: Another look
Let’s compare our results from table 1 to data from STATS. You can find STATS zone rating data online at Sports Illustrated. I’m using the numbers from the first calculation and not using the separate treatment for balls in and out of zone because I want to compare apples to apples here.
There are some big disagreements on a few players, like Edgar Renteria, Jack Wilson, and Hanley Ramirez, but the two zone ratings agree on most. The correlation coefficient between the two ratings is .827. STATS is using larger zones, showing more opportunities for every player. I don’t know what the BIS zones look like, but I can show you what the STATS zones look like here. This is not a big deal, using a slightly larger or smaller zone doesn’t make one measure better or worse than another, what is important is that is used consistently for all players.
What troubles me the most is not the opportunities, but the plays made column. A play made, whether in or out of zone, should be credited every time a player fields a groundball and records an out. Line drives and popups should not be included. Double plays started should count just the same as other outs, and when a player is the middle man in a double play, that should not affect his zone rating at all.
STATS originally counted double plays as two outs (for one opportunity) in zone rating, but changed their approach years ago and recalculated the original results. I don’t know why STATS is consistently counting more plays made than BIS, or what those plays are. Perhaps BIS is not counting all the plays they should?
I decided to delve into Retrosheet to find out what is going on. I looked at all 377 assists and 253 putouts that Orlando Cabrera was credited with in 2006, detailed what he did to earn every one of those, and tried to see how many plays he should have been credited with. An added bonus is that doing so with the OC allows me to revisit every Angels game from the 2006 season. Here are the results:
First, the Assists:
Most of the assists involve fielding a ground ball and throwing to first for an out. He should be credited with a play made when starting a ground double play, either a 6-4-3 or a 6-3 where he steps on the bag himself. We have 24 force plays on a flip to the second baseman, two runners thrown out at third, and three at home plate. There were two plays where multiple players were credited with assists though only one out was recorded. I don’t know how these were scored by BIS, but could possibly be considered as a play made, so we could be at 320 instead of 318.
One example came on 8/24 vs. the Red Sox and David Ortiz. Only one out was recorded, Ortiz at first, but the play was scored 6-4-3. This likely happened with an extreme shift, Cabrera probably got his glove on it and deflected it towards the second baseman who made the throw to first.
Now, the putouts:
Most putouts do not count as plays made in Zone Rating. You have popups, line drives, tags, and forceouts. Only two involved fielding a ground ball, Cabrera fielded a ground ball with a runner on first and stepped on second for the forceout. Strangely enough, both plays happened the same day, 4/7 vs. the Yankees. The first play ended an inning, but the next time was only the second out. Since Orlando did not turn the DP on that play, I wonder if Scioscia gave him a talking and ordered him to get the ball to the second baseman in that situation.
The most plays I could credit Cabrera with here is 322 with the two weird plays. BIS has 317, I wonder what they are missing. Could they be excluding the three ground balls where the out was made at home? I don’t think its right, but its not many plays and I could live with that. How does STATS come up with 347? This I don’t understand. I got that number from the CNNSI site, which has ‘CH’, total chances, and ZR. I multiplied CH by ZR to come up with plays made. Perhaps the numbers there do not mean what I think they mean.
I’ve been a big fan of Zone Rating since I first bought a STATS Baseball Scoreboard in 1989. After looking into the numbers though, I have more confidence in the new Zone Rating. I’ll look forward to checking these numbers daily on The Hardball Times this summer.