Saturday, 18 July 2009

Zambrano, Ramírez, Fontenot, Bradley

Messing around with the leverage index during yesterday's game and comparing it with the win probability reveals a disagreement. My method of multiplying the 'win value' of an event by the leverage does not have the same results as the win probability effect. Does that mean I'm doing it wrong? Maybe it does.

Player/Event/Inning       WPA        WV*LI
Zambrano/double/2nd 18% 13.9%
Ramirez/homer/3rd 11 4.9
Zimmerman/DP/1st 10 11.9*
Harris/BB/1st 9 7.6
Fontenot/double/2nd 9 8.6
Bradley/GDP/8th 5 7.6

I was a surprised at how the Ramirez homer scored so low in my method, not even as high as the Bradley double play. Curiously, though, it was the point at which the balance of the game shifted towards the Cubs permanently. It may be,in this case, that I'm rating double plays too highly. Are they the same as two outs, or the same as an out plus a caught stealing?
* I'm not sure I call this a double play. I've got two events, the Zimmerman strike out and Morgan being thrown out at third.

No comments: