MLB Defense Ratings, 2006 Season

[nanwynnfan]

FWIW, each season I compile defense ratings from a metric I've developed over the years. The basis of the metric is the assignment of degrees of difficulty for play execution at each position, the core element being the assist. The presumption is that the assist is generally the type of play that exhibits the athleticisms of the position player.

The putout is given a basic value for each position based on the observation that, of all the defensive entries on a scorecard, the PO is the "type" most prone to be routine over a long season. I realize that, like anyone who might have seen the game between the Bosox and the Yankees a few years ago, in which Pokey Reese hit the low retaining wall in foul territory up the LF line and folded like a card table - still making the catch in the stands - the PO was anything but routine. An inning or two later, Derek Jeter dove headfirst into the stands, sustaining facial abrasions and making another non-routine PO. It's not that POs count for nothing. They have a base value; but the assists are more heavily weighted.

The whole thing is part of a 1901-2006 study of MLB defense only, with changes of inputs due to changes in game dynamics.

I started the study based on the following suppositions:

1. defense has been historically the orphan child in the baseball ratings categories;

2. fans like to make comparisons of player skills, not only within a generation of play, but also among several generations; and I figured I'd try a study that makes the exercise feasible and credible;

3. not all fans have such intense sabermetric desire or dedication to subscibe to all the play-by-play [Retrosheet] granular breakdowns of performance, or the patience to study intensely mathematical studies of defense which often disagree in their results;

4. this attempts to base findings on the box score entries always there and published: PO, A, E, DP ... applied to the times, rules, equipment, ball, glove design, playing surface, etc.

Defense runs +/- are presented relative to league average performance at each position.

Example 1: Dustin Pedroia at 2B for Boston, gets a rating of +21.76 Defense Runs; but he played only 172 innings. Mark Loretta, playing 1,172 innings, gets a rating of +3.47. When all entries are made for Boston at 2B, the team rating us +5.81, weighted innings played determing team performance. Principal players are listed for each position; but I have not made entries for all players.

Base line for each position = +/- 0.00, which is the league average aggregated rating for each position. Ratings are position-specific.

It has been suggested there was a strikeout bias in my ratings, favoring low pitcher staff K teams over high K staff teams, based on fielding opportunity equivalence. The ratings provided here do not adjust, reflecting performance as originally calculated and evaluated. At the individual position level, such adjustments do not significantly effect defense runs, since the excess/shortfall in K's must be distributed among nine positions. On an aggregated team defense level they can alter ratings significantly depending on pitching-K variances of individual teams from league averages.

I've been told that my OF have high incidence of "relationship" with Range Factor, which essentially divides PO by games, which I do not do. However, given the regressions run to prove that point, I'll live with the findings and let the ratings stand on their own merits, being less concerned with their coincidence with another metric at particular positions [OF] than I am with results that promote discussion, debate or confirmation. Here goes.

Catchers
SEA: Johjima -1.63
1B
CLE: Broussard +4.65
SEA: Sexson +.98
2B
SEA: J. Lopez -4.21
WAS: Vidro -12.32
3B
SEA: Beltre +2.03
SS
SEA: Betancourt -3.43
LF
SEA: Ibanez +.72
CF
SEA: Reed -9.95
Suzuki +16.74
Bloomquist -4.37
RF
SEA: Suzuki +3.94
WAS: J. Guillen +21.61

MLB Defense Ratings, 2006 Season

[bedir than average]

Guys, i would ask that statements concerning nan's methodology go to the Saber board, but if you have comments concerning the particulars of the Seattle Mariners rankings and their implications for the 2007 club please do comment here.

One thing that I note is that either there is a flaw in the system which I will be talking to nan about over on the saber board, or the Mariners infield isn't all that good. Reputation at the left side is very high, tango's fan scouting report makes that obvious, but with the way the pitching staff is built (groundballers) a defensive metric that admits to being assisst dependent would ideally have a much better defensive infield for the groundballers to be successful.

Now the Outfield for 2007 looks like it could be one of the best defensive outfields out there, as long as Guillen is healthy and effective, even with the barely above average Ibanez in left field. The defensive replacements of Reed and Bloomquist one would think are much more suited to LF, and IIRC Reed's 2005 numbers were actually decent.

[Nimajneb]

Without taking the time just yet to dig into nan's methodology, a metric dependent upon assists would seem likely to underrate the '06 mariners defense relative to league averages...

[Chris]

Right, I don't understand how Betancourt is ranked so poorly?

[bedir than average]

Betancourt rated poorly in every non-scouting based metric I have seen in the past few months.