Batter's Box Interactive Magazine Batter's Box Interactive Magazine Batter's Box Interactive Magazine
When it comes to fielding metrics, there seem to be more and more on the market these days. The Raindrops is a Mets-focused blog. The author takes a look at various defensive metrics, using new Mets centrefielder Mike Cameron, and attempts to integrate them (and express the results in runs saved/cost versus average).

On the whole, he does a very good job. There are a few caveats to heed when you read through the article:

1) Win Shares is not based on play-by-play data; the others (to my knowledge) are. Ideally, another pbp metric would be substituted. Diamond Mind ratings are among the best out there and I think there's some kind of chart somewhere that can convert the grades (Excellent to Poor) to a runs scale.

2) It's a little unclear if the numbers given are related to range, or overall defence. UZR evaluates range and throwing separately, while Win Shares integrates them. David Pinto's system, judging by the name, concentrates on range. It's not clear whether throwing is or is not included in UZR data quoted.

3) The Win Shares adjustment compares Cameron to an average outfielder, but centrefielders in general get more win shares than corner outfielders.
Consensus is King | 6 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.
_Avkash - Wednesday, February 18 2004 @ 12:15 PM EST (#78580) #
http://theraindrops.weblogs.us
Thanks for the link. A few things:

1.) Rate2 is not based on play by play data. Davenport hasn't revealed the inner workings of his calculations, but I recall reading his DFT, which Rate2 is based on, compare an individuals defensive stats against league/team/ballpark/pitching staff averages. In this regard, its closer to win shares in methodology than the other two.

2.) While MGL gives consideration to DP rates and arm ratings, I've used only the pure range numbers on these charts.

3.) The only win shares data I could find (baseballgraphs.com) puts all outfield innings under one heading, regardless of which position. While I don't have the breakdown in front of me, I recall James's method properly crediting centerfielders against the corner guys.

4.) If you're good enough to start in center at the major league level, that is where you play. Corner guys move around the outfield, and even spend time in the infield, but true centerfielders spend all there time in center, with a few (small) exceptions.

Once I realized this, and recalling the inherited advantage centerfielders have in the win shares method w/o creating a seperate category, I didn't have a problem seperating by LF/CF/RF for the purposes of this article.
robertdudek - Wednesday, February 18 2004 @ 12:49 PM EST (#78581) #
I didn't realise that about Rate2, thanks Avkash. Maybe giving the p-b-p metrics more weight is a good idea (say a 40-40-10-10 weighting).
_S.K. - Wednesday, February 18 2004 @ 02:10 PM EST (#78582) #
For those too lazy to go through the charts, here's how the Jays made out:
Chris Woodward had a +3 rating, 13th out of 30 at SS. Most people will contintue to insist that he is terrible, of course.
Orlando Hudson had an outstanding +21 number at 2B, the best 2B in the majors and the second-best rating overall, to Cameron's +22.
Hinske had an awful -14 rating, 3rd-worst among 3Bs. Fingers crossed for this year...
Carlos had a rating of 0, exactly average at 1st.

In the outfield:
Vernon had a rating of +1, slightly above average.
In RF, Kielty had a slightly below-average rating of -4, while Reed Johnson surprised my with an awful -13 (only Tim Salmon was worse with a transcendental -21). Interestingly, Jose Cruz earned his GG by this metric, with a +18 number.
Cat didn't play enough, I suppose, but Shannon Stewart came out at a -2, showing that he isn't quite as bad as it would appear to the naked eye.

Great article, and I liked the methedology.
Craig B - Wednesday, February 18 2004 @ 02:40 PM EST (#78583) #
Chris Woodward had a +3 rating, 13th out of 30 at SS. Most people will contintue to insist that he is terrible, of course.

You bet I will. Well, not *terrible*. He's not good. It's not that I don't trust the metric, but I don't always trust the metric over my own observation.

Of course, I always see Woody on the fast turf at SkyDome, which probably biases me against him.

Vernon had a rating of +1, slightly above average.

And I don't believe this one either. I know Vernon isn't very good at coming in on balls in front of him, but he's so good going to his right, and so fantastic going back, I have to believe he's above average.

In the absence of proof, though, another person would be wiser to trust the numbers than to trust me.
Mike Green - Wednesday, February 18 2004 @ 02:53 PM EST (#78584) #
I'm not sure what happens when you separate out the "range" portion of UZR, but the chart does show that UZR has a large number of outliers (players with runs saved/cost using UZR more than 20 difference with any other method). The outliers are Polanco, Valentin, Karros, Kotsay, Dave Roberts, Eric Byrnes and Steve Finley. No other method had any outliers.

I'm sure that Steve Finley has lost range at age 38, but I doubt that he cost 53 runs more than the average centerfielder over a season. There just aren't that many plays. Similarily, I'm sure that Eric Karros is a modestly below average defensive first baseman, but I doubt that he cost 36 runs more than the average first baseman. The defence of first basemen is just not that significant in the game.
_S.K. - Wednesday, February 18 2004 @ 05:40 PM EST (#78585) #
Mike - I agree, that is very strange - especially when you consider that UZR, by methedology, seems like it would be the most reliable of the group.

Craig - I think Woodward looks worse in our subjective minds because of those terrible stretches he had last season where he seemed to be making an error every 5 chances. Overall, I think he has terrific range and I would have pegged him at slightly below average overall.
As for Vernon, I think it's too hard to judge outfield defence on TV (because you can't see breaks) for me to argue with you. But maybe he's just a good defensive CF who got some bad breaks last year, like a hitter who hits line drives right at people?
Consensus is King | 6 comments | Create New Account
The following comments are owned by whomever posted them. This site is not responsible for what they say.