Quote:
Originally Posted by Matt Arnold
A lot of the times the arguments come when talking about someone the old metrics love and the new ones hate, or vice versa, because people tend to get set in their ways. So when a new metric comes out that shows that Jeter is a poor defender at SS, the old school people who see highlight reel play after highlight reel play have grown up thinking one thing so it's hard to just shrug and "accept" that everything you knew to that point was a lie. Nobody complains when WAR says that Babe Ruth was the greatest ever player, because WAR hasn't changed anyone's view on him.
In time, people will learn to accept this, until 5/10/20 years down the road when people come out with something else that shatters some of what we know about WAR right now. Maybe we'll rediscover that steals are actually more valuable because some third order derivative of pitch values factoring in the catcher having to throw to the bag, and suddenly we're all going to have to rethink what we know right now about player value. We're already getting lots of new data about the effects of pitch framing that I think is still not all that common knowledge. I mean, when I look and see that Rralmuto has like the 4th or 5th best WAR in the NL, or whether Grandal really did deserve more MVP votes in 2016, there's still a lot of debate even in the advanced stats community there.
|
Is not likely that anything will "shatter" what we already know
https://www.beyondtheboxscore.com/20...ird-order-wins
Current versions of WAR already explain like 80-90 percent of the variation in wins. Sequencing (was it a walk then a home run or a home run then a walk) explains a significant portion of the unexplained part. If WAR correlated perfectly with actual wins, something would actually be wrong (unless we for to some next level brain [stuff] and we could know when a player was going to fail and succeed because of their clutchitude).
There will be advances, and tweaks, for sure. And WAR will get better. We'll learn more about how baseball works and how to measure it. But a real revolution in measuring players' performance is unlikely (in predicting or developing players is a whole nother story).
There's just not that much room left for shattering.