Thursday, May 03, 2012

How Badly Did Jim Tracy Blow It?

During the May 2nd Dodgers vs Rockies game, the Rockies were leading 5-3 with two outs in the top of the 9th inning (Dodgers batting). Matt Kemp was at the plate with Adam Kennedy at first base. The Dodgers were down to their last out and Matt Kemp represented the tying run at the plate. Even with a HR, the Rockies could do no worse in Kemp's at-bat than a tie ball game. The on deck batter was Dee Gordon, inserted in to the lineup in the latter part of the game. Dee Gordon represented the potential go-ahead run.

Worried about the hot hitting Matt Kemp, Rockies manager Jim Tracy decided to take his chances with the slender Dee Gordon by walking Matt Kemp intentionally. The strategy ended up backfiring badly on the Rockies as Dee Gordon hit a two run double, scoring both Adam Kennedy and Matt Kemp to tie the game. The Rockies went on to later win the game in the bottom of the 9th inning on a pinch hit three run HR by Jason Giambi, overcoming the managerial blunder by Jim Tracy.

Or was it a blunder? And if it was, how much of a blunder was it? I ran both scenarios, the pitch to Kemp and the walk Kemp through my simulator 100,000 times to see how the strategy effected the Rockies win expectancy. The simulator is good for these kinds of scenarios because it can take in to consideration the actual pitchers, batters and runners. A simple analysis would be to just look at the empirical data from past history to get an idea of the Rockies win expectancy in both scenarios. Below are my results along with the empirical results.

ScenarioDodgers Win ExpRockies Win Exp
Sim Pitch to Kemp5.64%94.36%
Sim IBB Kemp7.64%92.36%
Empirical Pitch to Kemp3.19%96.81%
Empirical IBB Kemp6.82%93.18%

So there you have it. The simulator says the Rockies lost around 2% of win expectancy by intentionally walking Matt Kemp to pitch to Dee Gordon. The empirical data says they lost just over 3.5% of win expectancy. Not exactly a game changing move, but definitely the WRONG move.

No comments: