The Origins of Modern Fielding Stats
Wisdom from Bill James in 1977
The Origins of Modern Fielding Stats
You might remember a video I made a few months ago about Dante Bichette’s awful fielding performance for the 1999 Colorado Rockies:
I thought about this video when I read this part of The Baseball Analyst’s review of the 1977 Bill James Baseball Abstract:
Take a look at all the clips in that Bichette video again. What you’ll discover is that Dante Bichette was almost always positioned incorrectly. In fact, because of his poor positioning and his leg injuries, Bichette was almost certainly one of the worst defensive outfielders in the history of baseball.
But, of course, that brings up an obvious question. How much did his poor fielding really hurt the Rockies?
Now, this is kind of a conundrum. Somebody asked me in the comment section the other day whether stolen bases should be charged to the pitcher or the catcher. This is a similar issue. The statistics we have for seasons before the year 2000 or so are so incomplete and so imperfect that it’s hard for us to figure out what was going on.
Baseball Reference estimates that Bichette’s defense was an incredible 34 runs worse than a replacement level player in 1999:
Fangraphs uses a similar methodology, and comes up with practically the same answer:
Note that Bichette only played in left field in 1999, though he did switch to right field when he went to Cincinnati in 2000. The -34 total zone rating is exactly the same as the -34 Rfield rating. This is because Rfield is total zone rating, as far as I can tell.
Now, I wish I could dive into this rating and how it was created. Unfortunately, I’ve never been able to find a specific explanation — something I covered a few months back in these two videos:
All snarkiness aside, there are a few things that cause me to doubt that Bichette was actually a full 34 run liability in left field in 1999:
Bichette only had a total of 268 chances in 1999 — and was credited with 238 putouts. Sure, there were theoretically hits to left field that an actually mobile outfielder probably would have caught. However, on what basis can we say there were 34 runs worth of those base hits?
Bichette’s fielding performance magically improved in 2000 when he played for the Cincinnati Reds. The big difference, as I explained in the video at the top of this article, is that he had Ken Griffey Jr. next to him in center field. Griffey got to those balls in right center field that Bichette couldn’t get to, while Darryl Hamilton and Edgard Clemente in Colorado weren’t able to get there. In other words — a good center fielder can make bad corner outfielders look good.
Colorado was a notorious hitter’s park in 1999 and in the surrounding years. There’s a really good chance that a lot of those balls hit to left field were impossible to get to for even the best outfielder.
So what does this have to do with the Bill James segment?
Well, the sad truth is that what James recommended is not the reality we have today. I can’t tell you anything about the number of 5 hole grounders that the best shortstop was able to get to last year.
The public facing sabermetric revolution was largely focused on hitting statistics. Fielding statistics, meanwhile, became the stuff of proprietary formulas and opaque boxes. Though Rob Neyer’s 14 year old article on FieldFX still makes me excited, the truth is that most of this data is obscure and remains difficult for the average fan to comprehend, let alone play around with.
Frankly, I think we could do better. Despite his many flaws, the one thing that was truly endearing about Bill James is the fact that the average person could easily understand his writing.
That — plus the fact that James never claimed that Bichette was responsible for giving up 34 runs in freaking left field.