Reading ahead in Bill James’ endlessly entertaining (and seemingly endless) series of articles about Win Shares and Loss Shares (on his website), we find that Ryne Sandberg, who won nine straight Gold Gloves, is never seen by Bill’s system as the best-fielding second baseman in the National League. In 1985, Sandberg comes in seventh ... even behind (gasp) Wally Backman. Bill:
Wally Backman had a horrific defensive reputation, and his numbers are generally very poor, but they’re good in 1985. Part of what is happening is this:
1) When things are poorly measured we lose specific information about them. We lose details.
2) When we don’t have detailed information about something, we tend to think about it in simplified terms, if not simplistic terms.
3) Because fielding statistics were not well designed, they don’t give us a lot of detailed, specific information.
4) This leaves us thinking about fielding in very general terms. We tend to assume that a player who is "good" or who gets good results in one season must be getting good results in the next season. It’s not true.
5) Fielding performance has the same season-to-season variation as anything else—the same as the things we can measure, like baserunning and pitching.
It’s not a big deal if Wally Backman rates OK defensively one year; it’s not evidence that our system has failed. It’s just something that happens—just like Carlos Pena playing at a Hall of Fame level in 2007.
Sandberg won nine straight Gold Gloves, 1984 through 1992. Again, Bill’s methodology doesn’t suggest that Sandberg deserved even one Gold Glove. Instead, he’s got ... well, a bunch of the other guys, including Glenn Hubbard, Tom Herr, Jose Oquendo, and Jose Lind.
Is Bill’s system “wrong” about Sandberg? Well, I don’t think Bill would claim that his system is perfect. I do think Bill would allow that in at least one of the seasons when Ryno comes out second or third in the league, he was actually the best fielder; no system’s perfect.
But Sandberg’s streak does suggest a dramatic inability of voters to subjectively determine who was really the best fielder in a particular season. Someone – might have been Craig Wright, but I just don’t know – made the interesting suggestion that Gold Glove voters might have done better back in the 1950s and ‘60s, when there were fewer teams and so the coaches and managers got to see everybody play more often. How are you supposed to see how good someone is when you see him play five or six times all season? Right. You can’t know. So you go on reputation, or maybe highlight plays. Sometimes that works really well; Alex Gordon and Andrelton Simmons come mind immediately. But oftentimes it doesn’t work well at all.
