More preliminary defense numbers
These use Sean Smith's system, Total Zone. He's done all players, all levels back to 2005, so we can start to see some trends. Important, because one year of defensive data is not generally all that useful, and minor league seasons are shorter than major league ones.
Here are notes on a bunch of guys you might care about. All the numbers are denominated in runs above average, so +10 means 10 runs above average. Keep in mind this is largely based on making plays, so while there is, in a sense, a penalty for making a bad throw that pulls the first baseman off the bag and doesn't get the runner, there's no additional penalty for corking it into right field and letting everybody take two extra bases. So error-heavy guys like Gamel should be adjusted downward.
- Iribarren: Great year at West Virginia in 2005 at 2B, when he was +14, otherwise average.
- Gwynn: Very impressive +19 in 2006, +4 in CF last year.
- Escobar: As noted a few days ago, he was +22 last year, which is either a massive outlier, or he improved a lot, since he was -13 in Brevard County in 07.
- Gamel: Always negative, including -12 for WV in 06 and -23 in 07 for Brevard County. He was only -2 last year, which we might attribute to Don Money, but of course we can't forget all those throwing errors.
- Cain: Of all the players I've looked at so far, he's the one whose TZ numbers most clearly point to an above-average defender. In RF, he was +27 (!) in '06 and +12 in '07 . The +12 is still kind of impressive since he played a fair amount of CF that year, limiting the amount of time in which he racked up the +12. In CF, he's consistently at or a couple runs above average in smaller samples.
- McGehee: Except for a flukish +10 in 2005, he's the very picture of average.
- Brad Nelson: Just as average as McGehee, only at more positions.
- Ryan Braun: to compare to Gamel, in 06 Braun was a total of -13 split between two levels. We don't have more than a half-season at any one level for Braun though.
- Bourgeouis: Because he's played so many different positions, we don't have a big sample for any of them. But from what we've got, looks like a bit above average at corner OF, little below at CF. In 2005, he spent more than a half-season at 2B and came in at +2.
- Chris Duffy: Consistently a few runs above average in CF.
- Taylor Green: In 06 in Helena, he was +6 at 2B. In 07 at WV, he was +5 at 3B. In 08 at 3B he was averae.
- Cole Gillespie: Consistently a few runs above average in the corners. Keep in mind this is a 'range' stat so there's no consideration of the player's arm.
- Caleb Gindl: It's only one season in low A, but Gindl was +13 in RF and +3 in CF last year. Worth watching.
- Brent Brewer: As with Gamel, keep in mind that he's worse than his 'range' number. That said, he's been consistently below average, including -9 in '07.
1 recs |
17 comments
Comments
Cain for '09!
And my mancrush on Jason Bourgeois is still intact.
Also, I wonder if Money’s work with Gamel at AAA prior to his call up coupled with his off season routine could move him up to an average defender at 3B. Either way, I’m guessing he’ll land at 1B at some point.
by Charlie Marlow on Jan 23, 2009 11:25 AM CST reply actions 0 recs
Wow, awesome work to both you and Sean
I just want to rephrase my point about samples in defensive stats. I’d consider a year of minor league defensive numbers to be about 200-300 at-bats, and we all know the fluctuations a player can have over that sample. So with Escobar being -13 and +22, it’s probably best to assume he’s about a +10 defender.
Inconsistencies like that tend to cause people to say that defensive stats are terrible and inconsistent. Here’s an analogy showing why that’s just not true.
Corey Hart had an .831 OPS in the first half and a .659 OPS in the second half. First half- about a .350 wOBA, second half, around .290. So if there were less total opportunities for plate appearances in a year, say 300, and Corey had these two wOBAs over two different years, you would see a stat like this:
Hitting runs above average, Corey Hart
Year 1: +5 runs
Year 2: -11 runs
See how that works out? So rather than conclude that the metric isn’t any good, try to get a bigger sample and look at the bigger picture. We’re not going to able to define the exact amount of runs saved or allowed, but we can get a good idea of defensive abilities.
Scored three times and detonated an indisputable in four visits to the batting box.
by Jordan M on Jan 23, 2009 12:07 PM CST reply actions 1 recs
Inconsistencies... cause people to say defensive stats are... inconsistent
That’s true.
Your analogy seems more of a single example of variance than an analogy. Logically its more supportive of the contrary to your argument.
I don’t see how increasing a sample size would overcome methodological flaws.
by ol Pete on Jan 25, 2009 11:26 AM CST up reply actions 0 recs
This is my first and final response to this. I'm done with debating defensive metrics and I don't want another long, drawn out discussion. Feel free to respond if you feel it's necessary.
Yes, it is an example of variance. When I first started reading about defensive stats, I was confused about how the same player could +10 runs and -8 runs in consecutive years, it didn’t seem to make sense. I think it might be the result of myself and other people equating “not saving or costing the team any runs” to “not saving or costing the team any runs, compared to an average defender”. There’s an important difference there.
I don’t think someone should look at Escobar being -13 and +22 in two years and thinking, “this stat must suck, if there is that much fluctuation, because two years can’t be that far apart”. I’m pointing out an example that shows that even a stable skill like overall hitting can be far above and below average within a sample similar to one year of fielding data.
And my point about sample and generalizing about defense stands. Jeff did a great job of explaining this in the post. He doesn’t say that every player with an above-average rating is an above-average defender, in fact, the only defender he concludes is almost certainly above average is LoCain.
If you have problems with the totalzone methodology, I would be interested to hear them. I’d suggest going to the Hardball Times archives and reading Sean’s series of articles about the system, when he tried to evaluate past players’ defense. I agree that sample wouldn’t overcome methodological flaws, but I’m fine with the methodology here.
I just don’t get how showing that hitting skill in relation to average can fluctuate within a sample similar to 1 year of defense to illustrate a point is supporting the argument that all defensive metrics suck.
Scored three times and detonated an indisputable in four visits to the batting box.
by Jordan M on Jan 25, 2009 11:51 AM CST up reply actions 1 recs
I’d suggest going to the Hardball Times archives and reading Sean’s series of articles about the system, when he tried to evaluate past players’ defense.
Acolyte.
by HRF on Jan 28, 2009 4:27 AM CST up reply actions 0 recs
Religious and anti-science. I can see the similarity.
by ol Pete on Jan 30, 2009 9:21 AM CST up reply actions 0 recs
I should add that the usual straw man arguments are present. That seems to be a standard technique.
by ol Pete on Jan 30, 2009 9:23 AM CST up reply actions 0 recs
Is there a drop in defensive quality between the minors and the majors? Obviously, it’s not like offensive or pitching statistics, but does the fact that the ball is hit harder at a major league level play a factor?
by tcyoung on Jan 23, 2009 12:25 PM CST reply actions 0 recs
I wonder
what else might play a factor. I’m assuming that major-league infields are kept in better shape, for example, and would be easier to turn ground balls into outs.
"I will agree that the attitude [at BCB] is ridiculous and they have done so much to instigate animosity and then block us from responding. Real mature!"
by roguejim on Jan 23, 2009 12:33 PM CST up reply actions 0 recs
Is the ball hit harder at the major league level?
I don’t know, but I’d be interested in knowing. My guess would be it’s not.
I don't specifically articulate my motives, because that wouldn't travel as well as a boo does.
by KLSnow on Jan 23, 2009 1:16 PM CST up reply actions 0 recs
I wonder
Part of me wants to say no, but another part of me wants to say that the ball is hit a little bit harder on average, or at least that the ball is hit hard more consistently.
I was just trying to give an example of something that could possibly play a factor. Roguejim expanded on the idea with another possible factor.
by tcyoung on Jan 23, 2009 2:43 PM CST up reply actions 0 recs
Interesting thought
On one hand you have the best of the best hitters around leading to the impression that balls would be hit hard a higher percentage of the time. On the other hand you also have the best of the best pitchers around throwing to those hitters, so there may be a canceling out effect. as rj also stated, the playing surface improvement might also cancel out some of the effect of harder hit balls, if anything I’m guessing there might be a slight drop-off in range.
by MadJimiBrewha on Jan 23, 2009 5:56 PM CST up reply actions 0 recs
Pitchers typically throw harder in the majors
So that may add to the speed of the ball coming off the bat. That would go along with the cliche, the pitcher supplies all the power.
by tcyoung on Jan 24, 2009 9:03 PM CST up reply actions 0 recs
Props on the defensive measure, but you’ll have to forgive me for at least asking how it rates gold glovers for runs above average. I don’t really have a frame of reference for the numbers above.
by ecocd on Jan 23, 2009 1:16 PM CST reply actions 0 recs
It probably depends
Are you talking Derek Jeter the Gold Glover, or are you talking Jimmy Rollins Gold Glover?
by Charlie Marlow on Jan 23, 2009 1:25 PM CST up reply actions 0 recs
Ozzie and Belanger were probably true-talent +25 to +30 run guys in their best seasons.
At each position, the best of the best are:
1B: +10 to 15 runs (Pujols seems to break the +10 "limit")
2B/3B: +15 to 20 runs
SS: + 25 runs
LF/RF: +15 runs
CF: +20 runs
Most “good fielders” are in the +5 to +8 range. +10 to +15 is pretty effin’ good.
Beyond the Boxscore // Calling BJ Upton lazy is lazy.
by Sky Kalkman on Jan 23, 2009 8:42 PM CST up reply actions 0 recs
maybe a random question...
but was wondering if you guys think Charlie Fermaint will be a ML player?
by Jamie in LA on Jan 23, 2009 1:39 PM CST reply actions 0 recs

by 
























