Identifying the Best Home Run Parks in Baseball History

Michael Schell

University Of North Carolina

Chapel Hill, NC

Home and away home run data is useful in identifying ballparks favorable for home runs.  By aggregating data across several consecutive years, random variation in estimated park effects can be reduced. A multiple changepoint method is used to identify when the park effect changes and to estimate its size in each time period (between changepoints), when it is modeled as constant.  The method identifies roughly 10-15 changepoints for each original 20th century franchise and park effects are estimated separately for left- and right-handed hitters.  Although changes are forced into the model only when the team moves into a new ballpark, most major stadium changes coincide with changepoints identified by the model.  Greenberg Gardens in Forbes Field, frequent fence changes by owner Charlie Finley in Kansas City (1955-67) and the effects on homers of the skewed dimensions of Memorial Coliseum in Los Angeles are among the effects that stand out.  By identifying the park effect, players’ home run totals can be adjusted, putting players on a “level playing field”.  Adjustment using the park effects reveals that Joe DiMaggio and Ted Williams each would have hit about 45 more career home runs than they did had they played in “average” parks.  On the other hand, 96 of New York’s Mel Ott’s 511 home runs, were the lucky by-product of playing in the home run-frie.