Friend of THT Tom Tango is running his 6th annual Fan Scouting Report. Hop on over and add your voice to the chorus. It only takes a few minutes. Bryan already told you about it yesterday, but I have a few things to say about the Fan Scouting Report. Plus, another reminder could never hurt.
There are certain realms where scouts will always be invaluable. Evaluating fielding is one of those areas, and it’s only partially due to the measurement errors inherent in all fielding metrics. Imagine the perfect fielding metric, taking into account the speed and trajectory of the batted ball, the positioning of the fielders, the handedness and speed of the batter, the batted-ball profile of the pitcher, and park. We’d still needs scouts to help us evaluate fielding. Why? Regression to the mean.
Regression to the mean depends on two things: one, the sample size, and two, the spread in skill among the MLB population. The smaller the sample size, the more you regress. The smaller the spread in skill, the more you regress.
The sample size involved with fielding is always going to be small, and that means that we have to regress the heck out of fielding metrics, even a perfect one, in order to estimate a fielder’s true skill. A full-time shortstop playing on a groundball heavy staff might see, at most, 500 chances to field a ball in a season (in some predefined “zone”). A centerfielder on a flyball heavy staff might have 400 chances. Pretty much all other fielders will have fewer chances than that. That’s smaller than the number of plate appearances a full-time hitter might gather in one year. So, on the basis of sample size alone, we’d have to regress fielding performance data more heavily than hitting performance data.
And that’s to say nothing about the spread in talent among fielders. I don’t know how big the spread in talent is, but I would venture to guess its much less than it is for hitting. A very good hitter might be something like 30 runs above average, whereas a very poor hitter might be 30 runs below average (over the course of a season). That’s a difference of 60 runs. I don’t think any of the advanced defensive metrics sees a difference of 60 runs between a very good and very poor fielder (maybe between the very best and the very worst, but certainly not between very good and very poor). So, the spread in fielding skills is probably smaller than it is for hitting, which means – again – we that have to regress more heavily for fielding than we have to for hitting.
All that regression, no matter how perfect the defensive metric is.
There’s another reason why the Fan Scouting Report is so important. From the instructions for filling out the ballots:
Try to judge “average” not as an average player at that position, but an average player at any position. If you think that Willie Bloomquist has an average arm, then mark him as average, regardless if you’ve seen him play 2B, SS, 3B, LF, or CF.
DO NOT CONSIDER THE POSITION THE PLAYER PLAYS!
In my experience, this is exactly how traditional scouts rate fielders. They don’t consider his position, but instead judge his range, first step, arm strength, etc. against all major league players. Why is that important? Because different positions require different skills. If we’re rating second baseman, footwork is more important than arm strength. If we’re rating right fielders, the opposite is true. This allows us to rank fielders by appropriately weighting the relevant skills at their position, but it also allows us to make predictions about how a fielder’s skills might translate from position to position.
So, our eyes can be a valuable aid to the statistics that we have available for fielding. And the Fan Scouting Report uses our eyes, the eyes that watch the Red Sox or A’s or Pirates or whoever over a hundred times a year, the eyes that have been watching baseball for five or 10 or 50 years. And, if enough of us participate, it will uses thousands of our eyes. So from one fan to another, please take a moment to contribute.