Baseball for the Thinking Fan

Login | Register | Feedback

You are here > Home > Primate Studies > Discussion
Primate Studies
— Where BTF's Members Investigate the Grand Old Game

Tuesday, March 06, 2001

A Defensive Stats Primer

Defense! Defense! Defense! Can’t tell the difference between a Zone Rating, a Range Factor, or a Max Factor? Hit the link to get clued in.

Baseball is made up of two things scoring runs, and preventing runs from scoring.?   The pitcher is all about keeping runs from scoring and the Designated Hitter   is all about scoring runs.? Any manager using a weak hitter at DH is clearly   unfamiliar with the purpose of the role.? Every other player has two parts to   his game ? his hitting and his fielding.

Offense has been studied thoroughly.? There are   many ways to evaluate offense: batting average (BA), on-base average (OBA),   slugging average (SLG), Runs Created (RC), Extrapolated Runs (xRun), Linear   Weights (LC).? The last three offensive summations do a really good job of describing   how runs are scored on a team level, and that is subsequently extended to the   individual players.? All of these methods are readily available around the Web,   with xRun?s progenitor being BB Primer?s own Jim Furtado.? You can read all   about it in the glossary.? You can calculate the approximate value of every   player, and essentially determine the league?s offensive MVP.

Defense is a bit trickier to measure.? Is it the   pitcher getting the hitter to hit into easy plays, or is there a big difference   in defensive players?? A quick glance at each team?s opponents? batting average   will tell you there is a limited amount of contribution a single defensive player   can make.? The important question is how much better is a defensive player than   the available alternatives, and ultimately his peers.

The common defensive statistics, unlike offensive   statistics, stink.? First there is ?Number of Gold Gloves Won.?? What do Gold   Gloves say about a player?? As I read Dan Szymborski write the other day, ?he   got the most votes.?? There is no objectivity to Gold Gloves whatsoever.? The   voters usually vote for the player with a good reputation, regardless of actual   skill level.? This really came to light in 1999 when Rafael Palmiero won the   award after playing just 28 games at first base for Texas.? Sometimes a good   fielder wins the award, sometimes not.? So using ?Gold Gloves won? is essentially   useless for determining who is a good fielder and who is not.?

Next is fielding percentage (F%).? It?s nice to   have a good fielding percentage, but it does little to tell you about how good   a fielder is.? Two shortstops get 100 balls hit in their general direction.?   Shortstop A fields 85 of them and makes no errors.? The other 15 are out of   his range.? His fielding percentage is 1.000.? Shortstop B fields all 100.?   He makes 5 errors (fielding or throwing).? His fielding percentage is 0.950.   ?The difference between the two is not that SS A is better; in fact, he?s worse,   but has a better percentage.? Pardon my blasphemy next: many of the ?great fielders?   were really just limited in range.? Larry Bowa was a fine shortstop and has   the highest fielding percentage for a career at short.? But Dave Concepcion   was much better.? He just could range further, turning grounders into outs that   Bowa was unable to reach.? Fielding percentage may be a good indicator of sure-handedness,   but it doesn?t do a good job at all for determining who the better fielder is.?

In the ?80s, Bill James came up with a statistic   called Range Factor (RF).? At first blush, this appears to be a good idea.?   It is calculated by dividing plays made by games played.? This provides the   number of outs a player made per game.? By comparing fielders to one another,   one could quickly see that Ozzie Smith was making 5.5 plays per game and Garry   Templeton was making 5.0 plays per game.? At 0.5 more plays a game for 160 games   meant 80 singles!? Wow!? And that?s how a legend is born.? Bill James is known   for giving rise to a new way to look at baseball: analytically, rather than   just nodding along to the clich?s.? RF is not a statistic to use if anything   else is available.? RF was modified to be calculated to plays per 9 innings.?   What RF misses is the critical aspect of chances.? The reason Ozzie made 0.5   more plays per game is because he had more balls hit to him.? Ozzie had a lot   more groundballs hit to him. That?s a big help.? The aspects of a fielder?s   pitching staff influence this statistic more than most. Range Factor, because   of its source and relative ?newness?, finds popularity in people just delving   into the atypical statistics of baseball.? To me, that?s bad; very, very bad.?   RF may have value over a career, but in today?s game, I think it is the worst.

So what do we do for a defensive statistic that   doesn?t stink?? We start scoring baseball games for the fielders in addition   to the hitters.? How?? We write down every ball that?s put into play, and where   it was hit on the field.? There is a company that does this:? STATS, Inc.

STATS, Inc. is a company that hires people to watch   baseball games.? These people write down every pitch.? They write down every   play.? This is where we get information about how hitters perform with runners   on second and two outs.? This is done to record what location every ball put   in play was hit.? This is where we get information about how many times a pitcher   throws to first base during each game.? Each ballpark has a grid with labeled   ?zones?.? When a ball is put in play, the STATS scorer records the pre-labeled   area of the field that the ball was hit to; how far the ball traveled before   being picked up; whether it was a fly ball, a line drive or a ground ball, and   an estimate of how hard the ball was hit.? The scorers do not make any judgment   with respect to whether or not a fielder should have made the play.? This is   critical because it removes the primary source of subjectivity in a player?s   evaluation. ?STATS usually has several people watching every game.? Multiple   game scores can be cross-checked to insure accurate ?zone? assignments for each   ball in play.? It?s really a strong network of data collection, but it is not   perfect.??

So you say, ?Dial, what does that have to do with   the price of peas in Peoria??? And I answer: it?s how we can see how well a   player does compared to his peers defensively.? The data accumulated by STATS   is reported (at STATS on AOL,, as a zone rating.? Every   player has one, and it is close to the fraction of balls a player converts into   outs divided by balls hit into a player?s zone.? A player gets a bonus for getting   to balls outside his zone.? For a good visual, consider the standard position   for each player:? where the shortstop stands, where the second baseman and centerfielder   stand.? For infielders, imagine a 20-foot circle around each player.? That?s   the area each player is expected to cover.? All shortstops are responsible for   the same zones.? Since they all are assigned the same zones, and they all stand   within a few feet of each other on the field, the percentage of balls they turn   into outs should be a mostly accurate description of how well they field.? Zone   Rating (ZR) is not a perfect measurement, and should only be described   as good.? It will undervalue defense because it doesn?t consider the ability   to get to pop-ups and turn double plays.? If you need to make a judgement on   how good a fielder is, try to use ZR at a minimum.

Unfortunately, there is no ZR data older than about   1985.? That means the Gold Gloves, the fielding percentages and range factors   for all the older players will have to suffice.? There are a few systems attempting   to measure fielding using play-by-play data, and they may improve on what we   presently have, but one must always recognize the shortcomings of these systems   and of the statistics gathered today.


Chris Dial Posted: March 06, 2001 at 05:00 AM | 3 comment(s) Login to Bookmark
  Related News:

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

   1. Alan Shank Posted: April 04, 2001 at 12:03 AM (#603618)
Mike Gimbel, a sabrmetrician who worked for Montreal and Boston under Dan Duquette, used Stats, Inc.'s play-by-play data in his "Defensive RPA" data. His stuff was the best I've ever seen, as it took extra-base hits into account and also was done on a park-by-park basis, so that differences in ballparks were taken into account.

He published a book called "Baseball Player Ratings" for several years, beginning in 1992 I believe. Actually, his stuff was on Jim Furtado's Web site for a while.
Alan Shank
   2. Chris Dial Posted: April 04, 2001 at 12:03 AM (#603619)
I haven't seen alot of Mike Gimbel's work. I've read some and exchanged a few ideas with him.
However, the first blush information I got left me unsettled with the way he did park factors,
as well as how he rated fielders. I'll try to get a hold of his books and learn more about his process.

   3. 185/456(GGC) Posted: March 22, 2005 at 02:57 AM (#1210477)

You must be Registered and Logged In to post comments.



<< Back to main

BBTF Partner

Dynasty League Baseball

Support BBTF


Thanks to
Infinite Yost (Voxter)
for his generous support.


You must be logged in to view your Bookmarks.


Page rendered in 0.3068 seconds
58 querie(s) executed