Normalizing MLB offense to stadiums

[Waterstpub87]

Working on a model with an input for Park Factor. Looking at the Park Factors, there seems to be a high level of variability from year to year in the same park. For example, in 2008 and 2009 AT&T Run effects were 1.045, and 1.05, favoring the hitter, and 2010-2013 .869, .737, .737, .942, favoring the defense. There are several other examples of similar behavior. Could this be a case of weather that year, pitchers pitching significantly better, or crowd effects?

Anyone else ever look at this

[evo34]

Working on a model with an input for Park Factor. Looking at the Park Factors, there seems to be a high level of variability from year to year in the same park. For example, in 2008 and 2009 AT&T Run effects were 1.045, and 1.05, favoring the hitter, and 2010-2013 .869, .737, .737, .942, favoring the defense. There are several other examples of similar behavior. Could this be a case of weather that year, pitchers pitching significantly better, or crowd effects?

Anyone else ever look at this

Some weather; mostly variance. Runs are variable enough that one season is not going to give you enough data to figure out a park factor.

[EXhoosier10]

run a regression with the last 1, 2, 3, 4, 5, etc years of data trying to predict the next years data and use whichever regression gives you the highest correlation or lowest RMSE

[Squared Box]

5 years of data is best.