Comments on: World Series Forecast

By: wwinston

wwinston — Thu, 05 Nov 2009 01:41:22 +0000

chance of winning is different each game because of home edge if nothing else. basically i use chess system adjusting for improvement in starting pitching due to shortening of rotation

By: Steve

Steve — Sat, 31 Oct 2009 16:05:22 +0000

So your logistic regression model gives you a probabilty of a team winning a game based on a bunch of factors (the X’s)? How many predictor variables do you have in that model? (I guess starting pitching is one variable?) I guess the point is it’s more accurate than just calculating a flat winnng probability that could be applied for all games in the World Series and using a binomila distribution to predict outcomes. Interesting stuff.

By: wayne winston

wayne winston — Mon, 26 Oct 2009 20:01:35 +0000

Rated teams based on a logistic regression model using all games. then played out world series 10,000 times factoring in home edge using @RISK.

By: Jeremy

Jeremy — Mon, 26 Oct 2009 16:56:19 +0000

Interesting stuff. Any background on how you came up with these numbers?