Better Probabilities

Looking at the specific pitchers starting should provide a better handle on the probabilities of the Cubs and White Sox winning game seven of the World Series. Instead of using the generic overall runs allowed by the teams in the Log5 calculation, I used runs allowed in games Kyle Hendricks and Corey Kluber started. In other words, what does the starter plus the bullpen allow in those games.

Hendricks started 30 games, and in those 30 games the Cubs allowed 91 runs, or 3.03 R/G. That’s in a league that allowed 4.48 runs per game. The Indians scored 4.83 R/G in a league that scored 4.52 runs per game. Log5 yields an expected runs per game by the Indians in a Hendricks start of 3.16.

Kluber started 32 games, and those 32 games the Indians allowed 122 runs, or 3.81 R/G. That’s in a league that allowed 4.47 runs per game. The Cubs scored 4.99 runs per game in a league that scored 4.44 runs per game. Log5 yields an expected runs per game by the Cubs in a Kluber start of 4.16.

Plug those numbers into the Pythagorean expectation and the Cubs probability of beating the Indians is .634. That’s very close to the .614 using just team runs scored and allowed.

There is one caveat, however. Kluber pitched in a game in which the Indians pen blew up. On July 3rd, Kluber gave up five runs in 3 1/3 innings, but the bullpen allowed 12 runs the rest of the game as Cleveland fell to Toronto 17-1. That game is a huge outlier, and Hendricks has nothing like that on his side of the ledger. Since the average number of runs allowed by the bullpen in a Kluber starts was one, I redid the calculation again, with six runs for that game instead of 17. Kluber’s runs per game drop to 3.47, and the probability of the Cubs winning drops to .584.

Neither of those probabilities is comforting for the Indians. The game tonight is not an even bet.

from baseballmusings.com http://ift.tt/2fEqSpM