Wednesday, November 11, 2009

Kouzmanoff - Defensive Record but no Gold Glove

Wednesday Novermber 11, 2009

Kouzmanoff was robbed.

The Gold Glove award is not a defensive award and it hasn't been for a very long time. They give these awards to anyone who hits well and plays decent defense for a team that gets press.

They gave the Gold Glove to Rafael Palmiero in a year played 28 games at 1B and another to Barry Bonds in a year he led the league in errors in LF.

Zimmerman hit .292 with 33 hr and 106 rbi. He does make alot of Web Gems, but he also blows alot of plays as evidenced by 17 errors. Only 4 players had MORE errors at 3B.

Kouzmanoff just makes every play. But he didn't hit 30+ home runs or play in Washington DC so he didn't get the press Zimmerman did.

Kouzmanoff's Web Gems weren't shown on SportsCenter because his games started after the 10pm EST SportsCenter started.

So the press didn't SEE him play. They simply don't know what he did.

And for those pushing UZR, did you know that it doesn't take into account ballpark or positioning? It doesn't take into account if you have someone playing next to you like Cabrera who makes a lot of plays deep in the hole. Actually it penalizes you if the guy playing next to you is a great player and gets to some balls in the bordering zone.

All it takes into account on range is whether a player made a play on a ball hit into a zone in which 50% of other players touched the ball in.

And I don't even want to go into the entire argument about the number of TC Zimmerman had versus how many Kouzmanoff had. I did two entire articles on that. Suffice to say the stats show it had much more to do with more IP by LHP and a higher ground ball percentage against those pitchers than with getting to balls farther away from the players original positioning.

Kinda sad when the supposed defensive award goes to a guy that is close to the worst at making errors rather than the guy who set a NL record for being the best.

13 comments:

  1. ThePad Father11/12/2009 12:04 PM

    I agree. Kouzmanoff deserved that award but between not being seen on SportsCenter and all the stat geeks that dont really understand fielding stats, he is not getting his due. Maybe if he played somewhere that he could hit 30 home runs?

    ReplyDelete
  2. You really have to wonder why so many so called "Padres" bloggers are trying to argue that Kouzmanoff didn't deserve the Gold Glove.

    ReplyDelete
  3. To your point about UZR not taking shortstops into play, Zimmerman's shortstop, Cristian Guzman, had a RngR of 2.0, while Cabrera was at -6.3.

    ReplyDelete
  4. Ray,

    That is one of the many problems with rating range by a zone in which 50% of the players at that position get to the ball hit there. It does nothing to determine how far players actually went to get balls.

    Cabrera made errors on some easy plays but he had one of the largest ranges in terms of how far he could go to get a ball of any shortstop in baseball.

    If you watched him play you will not even attempt to argue with that.

    some day we will have a Fielding F/X and then you will know exactly what a players range is. Until then all the metrics are flawed.

    ReplyDelete
  5. Web, I know I mentioned it a few times on my blog, but I'm not sure if you saw it.

    Anyway, UZR is not really what you're describing. It does not measure whether or not players get to a ball in one, generic zone -- that's RZR from Hardball Times.

    UZR looks a how a player did, compared against average, in many smaller zones. It also has a lot of adjustments for different factors, like parks and pitching staffs.

    I agree that all fielding stats are flawed, and there are still a lot of improvements to be made. But I don't really think it is fair to just throw away something like UZR, because it isn't perfect. Over multiple years, it does a pretty good job of measuring fielding.

    ReplyDelete
  6. Keep looking for new articles, but you seem to be gone. Whats up? Off season got you down?

    ReplyDelete
  7. Myron,

    Everything on UZR is readily available. BBTF has the best explanation, but there are others.

    Your range in UZR is defined by whether you got to more or less balls hit into those 64 zones they use than 50% of the other fielders in the league.

    Its 64 of the 78 zones that Retrosheet uses for its hit location diagram.

    It does not take into account where you were positioned or whether the ball is hit straight to the fielder, to the center of the zone, or to the far edge of a zone.

    Considering the Zones are an absolute minimum of 24 ft on a side, that is a pretty big area.

    If you play next to a guy with a huge range or that is a ball hog, then your range numbers in UZR fall, regardless of whether you were in position to make the play or not. In other words in UZR you are penalized by playing next to someone who is good. Please don't try to trot out UZR for Christian Guzman. We both have seen with our own eyes how superior Everth Cabrera's range is to Guzman's range.

    UZR also does not take into account pop flies and line drives whether fair or foul. If it is in the air, it doesn't count in UZR for infielders.

    Zimmerman only converted 18% of the balls hit in the air (pop fly or line drive) into outs. Kouzmanoff converted 23%.

    ROE and non-ROE fielding errors are not taken into account by zone.

    All ROE errors are treated as outs, so throwing errors where runners are on base due to the error are not initially taken into account. The fielder gets credit for making the out even if he threw the ball away.

    UZR does not take into account at all the prowess at turning double plays. Kouzmanoff was clearly superior in turning the double play. Nearly 50% better.

    If you have LHP throwing a greater % of innings then you are likely to have more balls hit your way as a 3B. If your pitchers induce a greater percentage of ground balls you will have more chances as a infielder. combine the two as was the case for Zimmerman then you get the benefit having many more ground balls hit to your side.

    I could go on but I have covered all this elsewhere.

    To say Kouzmanoff was not as good as Zimmerman because UZR says Zimmerman's range is better is faulty logic. UZR is faulty. Its lazy.

    Zimmerman MAY be better, but UZR and the current Fielding statistics do not show that conclusively.

    Committing 600% more errors does not indicate that Zimmerman is better.

    Getting to 45 more balls when he had 148 more hit his way does not indicate he is better.

    The fact that 3% LESS of his balls fielded resulted in outs does not indicate he was better.

    Look deeper. Don't use other peoples stats. The raw data is all available.

    ReplyDelete
  8. Web, you make a lot of good points, and I've enjoyed the discussion. I may have misinterpreted your interpretation of UZR ...

    Anyway, I don't think you can call it "lazy." Incomplete -- sure. But lazy? The reason it doesn't include pop ups and line drives is largely because, I believe, there is little variation in those skills between players -- especially line drives.

    This is of course debatable. But I don't think they are just left out for no reason. I have thought quite a bit about fielding stats, even messing with my own a few years ago, so I share a lot of your concerns.

    For instance, how about the speed of the runner that hit the batted ball. No fielding stat takes that into account, to my knowledge. A ground ball in the same exact spot, at the same exact speed by Ichiro is a lot tougher to turn into an out than one by Manny, because of the difference in their times to first.

    I don't think fielding stats will ever be perfect, even when we get to a fielding f/x system. They will always be evolving. I think Zimmerman is better than Kouz because he's better by the numbers *and* he's better by scouting. Again, I don't *know* that he's better. But with guidance from the info available, I think he is.

    The only real evidence that Kouzmanoff is better is his fielding percentage. And even if we wanted to include air balls, double play ratings, etc, we'd have to do a lot of adjustments for context. That is why I prefer to look at stats like UZR or RZR or +/- (they already do that, to a large degree), while acknowledging they are incomplete.

    ReplyDelete
  9. To the point of Cabrera versus Guzman, one of the tenants of Moneyball was the idea that the eyes tend to see what the eyes want to see. It looks like Cabrera has great range, but it also looked like Khalil had great range when he was diving all over the infield. Most advanced defensive metrics now say that Khalil was average at best, and most advanced defensive metrics say that Cabrera was near the bottom of the league this year.

    That, and it's hard for the eyes to have enough information to make a decision. I don't know about you, but I didn't watch nearly enough Nationals games this year to speak on how Guzman's defense looked.

    ReplyDelete
  10. Ray,

    I saw 6 Nationals games live and another 21 online.

    I do see a lot more of Cabrera than Guzman, so its possible that Guzman had games where he went further to catch a ball than I have seen that Cabrera has gone, but I think that statistically a 20+ game sample size is probably large enough to tell me if this is the case.

    There were no broadcast games I watched in which Guzman moved further from his initial positioning to catch a ball than Cabrera did.

    Next season Sportsvision will have a "Field F/X" available in more than just ATT Park in San Francisco and we will have definitive stats on range.

    Until then fielding metrics are guess work and trained eyes seem to work as well as zone based systems that cannot take into account positioning, reaction time, speed of ball off the bat and many other factors.

    I attempted to fund a "Fielding F/X" type project in 2008. I had a group of researchers using all available broadcasts to measure the distance that fielders ran to catch a ball. We completed the NL West for all fielders and outfielders in the NL before we run out of financing (About $25k out of my bank account).

    What we found blew the UZR rating out completely. They were sometimes close, but most were just completely wrong. Jody Gerut had 3 of the 5 longest runs to make a catch that season, but he was not in the to 5 in UZR or RngR.

    Please see what I have written here and in the forums you visit about the failings of defensive metrics. When you are done go read what Crasnick said about the same subject.
    http://sports.espn.go.com/mlb/hotstove09/columns/story?columnist=crasnick_jerry&id=4666623

    ReplyDelete
  11. Have you published these findings?

    -Ray

    ReplyDelete
  12. Also, among NL center fielders with 600 innings played, Gerut ranked 4th in RngR and 3rd in UZR/150 in 2008.

    -Ray

    ReplyDelete
  13. Ray,

    I always thought we would get around to restarting the project. I incorporated and bought a domain name and spent about $25,000 on collecting data before I realized that it was going to take 6 figures to do it for a whole season and I was not been able to find enough financial backing to continue.

    Then earlier this year I heard that Sportsvision was going to do something like it with multiple cameras in each ballpark (from what I understand they have been doing it under contract to the San Francisco Giants for the last 2 seasons), which would be more accurate than what we were doing plotting the information from broadcasts.

    That pretty much killed any opportunity to earn income from what we were doing, so I dropped it.

    I have shared some of the information with scouts I know and blogged about some of it, but that is as far as it has gone.

    Since its from the 2008 season its dated now.

    ReplyDelete