Originally Posted by BuckeyeRedleg
I have no problem with that. It just seems UZR is too big of a component of WAR.
I really don't understand how you arrive this perception. On what basis do you make such an assessment? Heck, how are you even calculating the % that UZR comprises? You can't just look at a guy's UZR and divided it by his WAR. Since UZR is adjusted positionally and has an expected distribution centered around replacement level, whereas offense has an expected distribution that is always positive (but which is centered at a difference place for each position), the math just isn't that simple. You can't go from 20 UZR and 4.0 WAR = 50% of WAR is UZR -- or any similar such calculation. You have to look at the range of potential values.
It's not like anybody just made up some rule that defense had to be worth X% of overall performance. It just is, based on observation and knowledge about how much certain kinds of plays affect how many runs are scored.
In terms of the calculation, I think it's helpful to think of it in terms of opportunity to produce -- all positive values. I think of it like this:
- The game is 50% producing runs and 50% preventing them.
- So, on the "how many runs is X player responsible for" scale, a team's position player gets up to 5 opportunity points. At the player level, that 5 point opportunity may be a bit more or less depending on where they hit in the lineup, but it's all pretty close.
- On the prevention side, the credit gets split between the pitchers and defenders. For the sake of argument, let's say run prevention is 80% pitching. That means position players get up to 1 opportunity point for defense, on average.
- Now, since defensive positions vary in their amount and value of opportunities, they range reliably from position to position. A DH caps out at 0. A 1B may be at 0.5, while a SS or CF is at 2. At the team level, they average out to 1.
So the way WAR works is it sets a baseline of performance. It treats all hitters the same, giving them a standard baseline of 0.5 points of performance. A few guys will manage to come in below the baseline, most will be above, some will be way above. Same goes on defense, except the baseline varies.
In practice, that means you have everything from a DH who earns 3.5 of the 5 possible offensive points but 0 defensive points. Compare that to a CF who earns just 1.5 offensive points but also earns 1.5 defensive points. If you tried to the math on that, you'd say that UZR accounted for 50% of the CF's value and it would look much too influential. But you have to separate out the realized performance from the range of possible ways a guy can contribute when assessing whether or not the scale of UZR is appropriate.
I realize that was pretty abstract and probably makes no sense to anybody but me. But I think people have a lot of trouble placing defense and offense in the same context since the talent distributions are so different and offense is all reported from a standard baseline whereas defense is reported positionally.