Thursday, March 7, 2013

Batter/Pitcher Righty/Lefty splits begining with 1950

See previous posts on problems with the accuracy of aggregate splits data.  Play-by-play data is pretty darn accurate, box score data is not.  The percentage of annual data derived from play-by-play data reaches an acceptable level in 1950. has 82% of splits data in 1950 from play-by-play, rather than aggregate data from box scores. has about the same accuracy but its data is more difficult to copy into a spreadsheet so I'm going with baseball-reference.

I won't be posting one year at a time but I thought I'd share the results of the first year that I tried.  Righty batter facing a righty pitcher is the base and below are the percentage differences.

Split                     HR/AB  BAbip   BA  OBP  SLG  OPS
vs RHP as RHB base base base base base base
vs RHP as LHB -10.65% 5.86% 7.75% 7.42% 4.80% 6.15%
vs LHP as RHB -0.14% 0.00% 3.49% 2.37% 2.53% 2.60%
vs LHP as LHB -15.23% 0.00% 0.78% 2.08% -2.27% 0.00%

RvR - best at hitting home runs.  Lefty batters against righty pitchers sweep everything else.

The zero numbers are just too small to display in percent form but the underlying difference is there and will be used in future calculations.

Here's the current plan.

1. Enter data for seasons 1951 through at least 1970.  This is a pain so I want to limit the work to the era of the most interest.

2. Plot the percent differences on graphs and look for patterns.

3. Apply the averages to a few star batters of the 1950s and 1960s, probably Willie Mays, Hank Aaron, Frank Robinson and Ernie Banks.  They are all home run hitting righties.  I may try it for two lefties for whom we have only partial data since their careers started well before 1950:  Ted Williams and Stan Musial.

One objective is to get an idea of whether they could have done better if they had switch hit.  The basic methodology will be to apply the percentage differences for all MLB batters during that era to how they hit against righty pitchers.

Let's say that a righty batter had a .300 batting average (BA) against lefty pitchers and .280 against righty pitchers and that all lefty batters hit six percent better against righty pitchers.  We can assume that the BA of a great righty batter would be at least six percent better against righty pitchers if that batter had batted lefty against them.  That would make his BA against righty pitchers about .297 and that would be applied to the 72% or so of his at bats against righty pitchers to figure a new season BA.  As indicated in previous posts, many right handed players may have batted lefty by choice.  In other words, anyone can learn learn to bat lefty.

4. Look at Mickey Mantle's numbers in much more detail using year to year data, not averages for all the years he played.  Also, use different home run rates home and road since Yankee Stadium during Mantle's era was very difficult for righty batters to hit home runs.  Father knows best may not apply.

5. Take a summary look at other notable switch hitters: Eddie Murray, Chipper Jones, ...  I'll think about this.

No comments: