Situational Home Run Hitting

Situational Effects

As some of you might know, one of my interests over the years has been in situational effects — these effects such as home vs away, opposite vs same arm pitching, clutch vs non-clutch situations, are often subtle and hard to interpret.  To illustrate situational effects in baseball, I thought I would take a “situational” historical view of home run hitting.  We all know about the general effects  (home run rates are currently on the rise, for example), but how home runs are hit in various situations is less known.

Data

For this study, I used Retrosheet play-by-play data to collect situational effects for all home runs hit in the seasons 1980 through 2015 — in these 36 seasons there were 154,294 home runs hit.

Average Runs Value

The benefit of a home run is commonly measured by runs value found by using the runs expectancy matrix.  In Chapter 5 of Analyzing Baseball Data With R, we found the average runs value of a home run was approximately 1.39.  How has this average runs value changed over the years?

hr_situation_1.png

What is interesting is the volatility of this average measure over season.  It is not easy to see a basic trend, but it appears that the average runs value of home runs has been decreasing in recent seasons.

Solo Shots?

If the average runs value is decreasing, that might mean that there is a greater percentage of solo home runs.  This motivates graphing the percentage of solo home runs over our time period.

hr_situation_2.png

Like the first graph, there is much variation in this percentage over seasons, but the percentage of solo home runs is indeed increasing in recent years.  That seems to imply that teams are less effective in setting the table; that is, getting runners on base.

Grand Slams?

If solo shots are increasing, that might suggest that grand slams (home runs with the bases loaded) are less prevalent.

hr_situation3.png

The graph shows that the percentage of home runs that were grand slams was high in the mid 90’s (and reached a high in the 2000 season), but this percentage has been consistently low between 2011-2015.

There is No Place Like Home?

We know that this is home field effect in scoring runs — the home team will generally score more runs at home than on the road.  So one would expect there to be a home bias in hitting home runs.  To check this, I graph the percentage of home runs hit by the home team across seasons.  I have added a line at 50% so we can quickly see if there is indeed an advantage.

hr_situation4.png

There are three interesting patterns in this graph:

  • There was not a home advantage to hitting home runs between 1980 and 2000 — in fact the smooth lies completely below the 50% line.
  • There was a boost in the home advantage peaking about 2008.
  • The percentage of home runs hit at home has been decreasing in recent seasons.

Count Effects?

It would seem that batters would hit a high percentage of their home runs in so-called batter counts (like 2-0 and 3-1) where the pitcher is more likely to throw a ball in the middle of the strike zone.  Likewise, I would anticipate batters to hit a small percentage of home runs in pitcher counts (like 0-2 and 1-2).  I’m actually more interested in seeing how these percentages have changed over seasons. (By the way, the careful reader might note that I’m limiting this analysis to seasons 1988 and later.  The percentages of home runs hit in batter and pitcher counts in the Retrosheet data were much smaller in seasons before 1988, suggesting the data is not complete for those earlier seasons.)

hr_situation5.png

In the 1990’s, the percentages of home runs hit in batter and pitcher counts were about the same (25%) — this means that about 50% of the home runs were hit in neutral counts (0-0, 2-1, etc.)  But these two percentages have diverged greatly in the last 20 seasons.  Hitters now are hitting a greater percentage of home runs in pitcher counts.  There likely is a good explanation for this that could be gleaned from further study.  (For example, I did not account for the number of pitches in batter, pitcher, and neutral counts — these may have also changed over the years.)

Takeaways from this work

One thing that is obvious from this study is that situational effects show much variability across seasons — you can’t make much sense of these by looking at a single season.  Generally, situational effects are a different animal from main effects (like summary performance measures for batters and pitchers), and it is easier to be misled by some of these values.  But a historical study does lead to finding interesting patterns.