Traditionally, statistics like wOBA (weighted on-base average) have been calculated using league averages. While building the baseballDBR package, I thought it would be interesting to group the American and National when making wOBA calculations. In theory, there should be parity across the two leagues, but that is not always the case.
In order to calculate wOBA values for each league, the baseballDBR package uses a ported version of Tom Tango’s SQL incantation to calculate wOBA using the Baseball Databank. While Tango admits, this calculation is not perfect, it normally has a plus/minus of less than one one-thousandth of a percent compared to Fangraphs’ values.
Gathering wOBA Modifiers by League
Plot Leauge wOBA Vales by Year
The plot shows the modern parity we expected. It also shows the affects of the “dead ball era” prior to 1920. However, what is interesting is the increase in league wOBA between 1920 and 1930 in the American League It should be mentioned, the stat “league wOBA” here is the average OBP, or on-base percentage for each league.
Using OBP to Find Outliers
There were obviously players in the American League between 1920 and 1930 that were blowing the curve, performing well above average. We can drill down deeper to find out exactly who those players were. Note, since the league wOBA represents a league average OBP, we will use OBP instead of wOBA to find our outliers.
There you have it, George Herman Ruth, blowing the curve once again!
To leave a comment for the author, please follow the link and comment on their blog: Data Science Riot!.