Quote:
Originally Posted by nat
Don't listen to Nick about B-R, it's a very useful site. Philosophically divergent from Fangraphs on a few issues, but that's all. bWAR (the one on baseball-reference) and fWAR (the one on fangraphs) are useful for different things and in different ways (pretty much as with any two stats).
We know that fielding is the roughest part of the WAR formulae (and the hardest thing to measure in any case). Nick is right that it gets less accurate the further back in time that you go. The folks who put the various WARs together decided to use the best defensive measurements available for each season. Since our measurements improve over time, this means that the numbers that go into the WAR formula for a player in 2016 aren't quite the same as those that go in for someone who played in 1966. They could have used the same measurements all the way through, but at the cost of making evaluations of modern players less accurate.
This means that we know that there are some errors in our evaluations of older players, we just don't know who is being affected by the error, nor precisely how significant the error is. People in 1966 just weren't recording enough data for us to be able to tell these things. Clemente may have been better than dWAR gives him credit for, but there's no way to tell, and there's REALLY no way to tell how much better.
|
I'm just giving him guff about BR dating back to that Jim Kaat argument.
It is useful, I just don't think much of their version of WAR.
you bring up a good point about older (and even present day) defensive numbers, they could equally overestimate as underestimate defensive value, people tend to lean towards their perception of a player, but that's usually not a good idea.