May 25, 2017
You Gotta Love a Slugfest
Recently on Effectively Wild, listener Andrew emailed: “How many runs need to be scored in a game in order for it to be considered a slugfest? Is it strictly a runs thing? Do a certain amount of home runs need to be hit? Do both teams need to be doing the slugging?”
Hosts Ben Lindbergh and Jeff Sullivan kicked around the definition of a slugfest. They didn’t come up with a firm definition, though Lindbergh thought that both teams had to score double-digit runs.
The following episode began with a follow-up, as Baseball Prospectus author Meg Rowley reported that there were 1,269 games since 1913 in which both teams scored 10 or more runs, 84 of which had no homers (last occurrence: June 22, 2008, Royals 11, Giants 10), 424 had five or more homers, and 11 had 10 or more homers. Since Rowley’s search, one more game met her criterion: Orioles 13, Tigers 11 on May 16, with seven home runs.
An obscure metric that can be measured over decades ... right up my alley!
At this point, rather than wait until the end of the article, I’d like to acknowledge Andrew for the great question, Lindbergh and Sullivan for selecting and answering it, and Rowley for sharing her very clever Play Index code with me.
So let’s talk slugfests. Lindbergh and Sullivan loosely defined a slugfest as a game in which each team scored 10 or more runs, and Rowley’s Play Index search used that definition, so I will too. By that measure, there were 10 slugfests last season. There were 10 in 2015, too. There were eight in 2014, six in 2013, 12 in 2012, and just five in 2011. Those five slugfests in 2011 represent the low-water mark in the 30-team era and the fewest slugfests since 1989, when there were four.
We’re certainly not living in a Golden Age of Slugfests. There were 41 slugfests in 1999, the most in history. There were 40 in 2000, the second-most. There have been 15 seasons in which there were more than 20 slugfests, and eight of them occurred between 1993 and 2004. That timeframe, roughly coinciding with the Steroid Era, would suggest that more homers equal more slugfests.
That conclusion would be wrong for two reasons.
First, there’s a difference in games played. From 1901 to 1960, there were two leagues of 16 teams each, and each team played (assuming no cancelations or replays of ties) 154 games per year. That meant there were 2,464 team games per season. Now, there are 30 teams playing 162 games each. That’s 4,860. So since the addition of the Diamondbacks and Rays in 1998, there have been 97 percent more baseball games per season than there were through 1960.
So we can’t really look at 1951, when there were 14 slugfests, and 2005, when there were also 14, and declare “same thing” because the slugfests in 2005 occurred over many more games. There are two ways to correct for this. One is to simply look at per-game averages. That works for events that frequently occur. If you go to Baseball-Reference, for example, you can get per-game averages for many hitting, pitching, and fielding metrics. Using per-game averages automatically corrects for expansion, games in the schedule, shortened seasons, and cancelations/replays.
But per-game averages don’t work well for relatively rare events like slugfests. Last season, the average game featured 8.96 runs, 16.06 strikeouts, 2.31 homers, and ... 0.004 slugfests. That makes no intuitive sense. So I’m going to present slugfests per year, as if each year comprised 30 teams playing 162 games. The math involved is pretty straightforward:
Adjusted slugfests = Actual slugfests x 4860 (30 teams x 162 games per team) games / actual team games played
So in 2005, when there were 14 slugfests, the adjusted figure is 14 x 4860 / 4862 team games played in 2005 = 13.99. In 1951, when there were also 14 slugfests, the adjusted figure is 14 x 4860 / 2478 team games played in 1951 = 27.46, almost twice 2005’s adjusted figure.
Here are adjusted slugfests per season, from 1920 to 2017:
As you can see, there were two peaks in slugfests: The Lively Ball Era in the 1920s and 1930s, and the Steroid Era from the early 1990s to the mid-2000s. That big spike is 1930, when there were 34 unadjusted slugfests, the third-most in history, despite it being a 16-team, 154-game season. For those unfamiliar with early 20th century baseball history, 1930 was a remarkable year, sort of the opposite of 1968’s Year of the Pitcher (but without a catchy sobriquet).
Here are some of the things that happened in 1930:
Here’s what 1930 was not, though: A homer-rich environment. That’s the second reason we can’t simply say homers equal slugfests. As I pointed out in this article, relatively few players hit home runs back in the 1930s. Only 19 batters hit 20 or more homers in 1930, compared to 111 in 2016. The average team hit 0.63 home runs per game in 1930; we’re running at 1.22, which would be an all-time high, so far this year.
So it’s not homers that drive slugfests. The correlation between adjusted slugfests and homers per game is -0.18, indicating no statistically significant relationship. Then if not homers, what’s behind slugfests? It’s an even simpler answer: Runs. The chart below is the same one as above, but I added a yellow line that’s average runs scored per team per game, scaled along the right axis.
See that? The most runs scored per game in modern baseball history was 5.55 in 1930, followed by 5.19 in 1936 and 1929. Those three years rank first, fifth, and third, all time, in adjusted slugfests. The fewest slugfests, outside of the World War II-depleted 1943 season, were in the aforementioned Year of the Pitcher. The correlation between runs per game and adjusted slugfests is 0.82. That’s an exceptionally strong correlation. It means that you can describe most of the year-to-year variance in slugfest frequency by run environment.
And that’s why the slugfest upturn of late is just that—an upturn, not a challenge to prior records. Home runs are leaving parks at a record clip, but run scoring remains pretty sedate. Scoring per game so far in 2017 is 4.57 runs per team, which ranks only 42nd in the 98 seasons since 1920. Scoring’s up—by more than a run per game since 2014—but not enough drive an unusually high level of slugfests.
Finally, Lindbergh and Sullivan wondered what percentage of games are slugfests. The answer: About 0.7 percent since 1920—comprising of 0.8 percent of games from 1920 to 2005, and 0.5 percent since the implementation of the Joint Drug Agreement in 2006.