The Hard Part

My Fielding Linear Weights metric is flawed.  The unfortunate part about the flaws on the metric are that they are hard to fix.  By hard, I mean I don’t exactly understand how to fix them.  Bill James sums it up best in my mind.  He said something along the lines of the fielding stats from a poor team are not that much different from a good team.  I believe I can relate to that on a player level.  My stat does not understand how to separate performances from good and bad.  It usually just comes up with random numbers in which one should have little confidence in.  The numbers seem to close together, bad fielders rate the same as good fielders.  That was my goal to fix, to determine good fielders from bad fielders before I even began work on the metric.  I don’t know if my formula just wasn’t in depth enough or not.  It seems to me that’s part of the problem.  Not enough weights, not enough in the formula to make it different. 

Then the flawed nature didn’t stop there however.  I put too much stock into Range Factor being a great metric.  Without adjustments for groundball staffs or fly ball staffs, plus more issues that need to be adjusted for on Range Factor.

 

The reason we need to get this fixed is because as of now for historical Total Runs counts for players I do not have and will never have Defensive Runs Saved data.  So, I have to rely on something nobody can rely on.  Fielding Runs.  Fielding Runs, being criticized by a lot of people is not too accurate.  My goal is simply to create a metric that has more accuracy then Fielding Runs.  My estimated Runs Created formula took me like ten minutes to create and I find it pretty good indication of how good a player batted in a season.  But fielding, with Fielding Runs is shaky at best.  So that’s our goal create a reliable Fielding Linear Weights metric then apply to make our Total Runs metric the best out there. 

Thanks for reading.