WPA Curiosities

And that, right there in a nutshell, is the weakness of WPA.  When the new theory goes against common sense, maybe the theory is wrong.

I’d say it differently.  There’s nothing wrong at all with the theory; it is what it is and it’s a great tool for certain uses.  But we should be aware of its limitations and only use it appropriately.

Well frankly I think that Zito’s 8 innings of 1 run ball is more valuable than two solo home runs.  As for Street finishing higher in WPA, Thomas may have had a “better” game that more often would be more valuable.  But thats different from a evaluation that says that Street’s inning provided more value to his team in the context in which it came.  That may seem odd and couterintuitive, but all that WPA is telling us, is that based on the leverage of the moments that each one (Thomas and Street) had, Street’s moments were more valuable - regardless of who’s game was “better.”

What I’m saying is, this could just mean the weights are wrong.  The numbers need jiggled.  Hey, I play with a lot of formulas, and you get a feel for these things when they don’t work out.  There are the times when you need to trust the formula, and times when the formula needs fixing.  My gut tells me this formula needs fixing.  Not so much because Zito rates ahead for the game, because the value of eight shutout innings is tremendous.  But because other hits ranked higher.

Formulas and theory are two different things.

I can assure you that the formulas are valid.  The output of the spreadsheet is in line with historical outcomes, which you can find here.

You just have to recognize WPA for what it is.  Frank Thomas may not have hit his home runs in “crucial situations” at the time, but the fact that he did hit them affected the LI numbers for the rest of the game.  There is something to think about here, a weakness of WPA, which is that when you look at the entirety of a game after the fact, each run scored ends up being just as important as any other.