Fielding Statistics Glossary

Statistic Description Formula
Inn Innings Played at position
Note: only available from 2000 onwards (see Note #1 below)		 n/a
PO Putouts n/a
A Assists n/a
E Errors n/a
DP Double Plays n/a
PB Passed Balls - catchers only n/a
FA Fielding Average - the number of chances accepted (PO+A) divided by the same plus total Errors (PO+A)/(PO+A+E)
UPO Unassisted Putouts - 1B only: this is an estimated team fielding stat, created by Bill James and outlined in Win Shares. It is an attempt to estimate the number of unassisted putouts the 1B for a team made during a season. In theory, this is a much more useful number than total putouts, as most of those are made by simply catching the ball while standing on the bag. UPO are estimated by adding the Assist totals of the team's 2B, 3B, SS, and P, and then multiplying this total by 0.84. This amount is then subtracted from the total PO of the 1B. Ainf=A(2B)+A(3B)+A(SS)+A(P)
UPO=PO(1B)-0.84*Ainf
UPO+A Unassisted Putouts plus Assists- 1B only: When adding estimated UPO together with Assist totals for a 1B, this gives a good estimate at the total number of plays made by the 1B. This value can then be used to get a good idea of the 1B range. If only assists are looked at, this discounts those 1B who choose to make plays themselves, instead of tossing to the pitcher. Read the "Garvey vs. Buckner" essay in Win Shares for more detail, it is quite convincing. Ainf=A(2B)+A(3B)+A(SS)+A(P)
UPO=PO(1B)-0.84*Ainf
RF Range Factor - the number of (Putouts + Assists) per game (or 9 innings played) ((PO+A)/Gm) or (9*(PO+A)/Inn)
lgRF league average Range Factor - the number of (Putouts + Assists) per game for the entire league at that position, using total Games Played at the position (issues) (PO+A)/Gm
RF+ adjusted Range Factor - a player's Range Factor normalized to the league average for that position; 100 is league average. Remember that a major league average defender is typically quite good. (i.e. an average 2B in MLB in 1950 was possibly one of the top ten or so 2B on the planet.) (also see Note #2 below) 100*(RF/lgRF)
Range An attempt to effectively measurement a fielder's true range.
For C,1B,2B,SS,3B,P; Range = Assists/Game.
For OF; Range = PutOuts/Game.
For 1B (team stats); Range = (UPO+Assists)/Game
I don't recommend using Range+ to rate the effectiveness of a 1B at the player level. Simply measuring a 1B's assists is not effective. This is why I have added the UPO stat under Team Fielding. I recommend using this for a more accurate measure a 1B's true defensive ability.
Note: Pete Palmer in Total Baseball defines Range in this way (except for team 1B). Palmer actually calculates a league-adjusted value for Range, a more useable value for comparisons. I have designated this as Range+. See below.		 C,1B,2B,SS,3B,P: A/Gm
OF: PO/Gm
Team 1B: (UPO+A)/Gm
lgRange league average Range - the Range for the entire league at that position, using total Games Played at the position (issues) (PO+A)/Gm
Range+ League-adjusted Range - a player's Range normalized to the league average for that position; 100 is average. A fielder with a Range+ of 105 gets to 5% more balls than the average fielder of his era.
Note: This is similar to the Total Baseball 'Range' rating, but not exactly the same. Total Baseball claims to use adjustments for playing behind L/R pitchers, etc., which I have not used. (also see Note #2 below)		 100*(Range/lgRange)
Arm Defined as 'Assists per Game', and valid for OF positions only. Not terribly useful on its own, as there is a large variation in lg average A/Gm depending on era. See Arm+ below for a more valid metric. A/Gm, for OF only
Arm+ This is an attempt to measure the quality of the throwing arm of an outfielder using Arm, as defined above, and normalized to the league average A/Gm for outfielders. Again, an Arm+ rating of 100 represents the league average. An outfielder with a Arm+ of 110 throws out 10% more baserunners than the average outfielder of his era. 100*((A/Gm)/lg(A/Gm)), for OF only

Note #1: Data for innings played at defensive positions by each player is only available from year 2000 onwards. As a result, I have used 'per Game' instead of 'per 9 innings' for all defensive metrics. I realize that using innings played would be a more accurate method, but that data is, unfortunately, not available to me for most of baseball history. This will not affect everyday starters too much, but it will make the analysis of 'defensive replacement' types practically meaningless.

Note #2: Range+ and RF+ values are of questionable value when analyzing the OF position. This is due to the fact that all OF positions all lumped together when calculating the league average Range or RF. There is not an even distribution of chances for all OF positions. Center Fielders typically get 25% more balls hit to them than a Right Fielder or Left Fielder. As a result, CF will look even better than they were using RF+ and Range+, while LF and RF will appear to be less effective than they actually were. I've thought about a few ways to weight the league averages to better capture an outfielder's true RF/Range above the league average at HIS primary position, but it gets messy. I have left it as-is for now.