The Supposedly Decreasing Importance of Strikeouts

Note: I have no idea if I’m the first to do this, but quite frankly I don’t care.

Let me start by apologizing for the Papelbon thing. It was a pretty stupid article, and I was basically just looking for something to write about. While I’m at it, I should probably apologize for the bFI thing–I thought that would come out better than it did–and the last part of the Pettitte thing–when a guy’s gone 28-6 against you, you tend to harbor some animosity towards him. With all that said, I feel like this is a pretty good one, even if it is rather brief. So, without further ado…

By now we’re all sick of hearing it. Strikeouts don’t matter anymore for hitters! They’ve lost their stigma¹! These crazy kids today don’t know about plate discipline! For the most part, these criticisms all seem to be saying the same general thing: Strikeouts (or the lack thereof) are no longer correlated to offensive success.

Well, I can’t speak for you, but I have really grown sick of these baseless assertions. Other writers have touched on the fact that there is virtually no correlation between strikeouts and offensive performance², but these are all within the past several years. What I wanted to prove was that there has never been a correlation between the two.

The methodology was pretty simple: Since wRC+ is the tell-all offensive statistic, I simply found the correlation, measured by R-squared³, between K% and wRC+ for every season from 2012 going back to 1913 (the first year that strikeouts were recorded for batters). I then graphed the resulting R-squared⁴ values by year for every year, of which there were 100.

And what, you ask, were the results?

Graph

“Well, golly, them folks was right!”, the reader might be inclined to say. Indeed, it would seem that–although the R-squared values have fluctuated heavily over the years–they are, overall, at a lower level than they once were. This would mean, of course, that strikeouts did matter more in the days of yore.

But wait! All hope is not lost! For you see, I purposefully excluded one key aspect of the graph in question: the labeling on the y-axis (i.e. the one upon which the R-squared values were measured, i.e. the vertical one). Put that back on, and what do we discover?

Graph2

For the entirety of baseball’s history, there have only been FOUR YEARS with an R-squared above .1. Remember, R-squared is on a 0 to 1 scale, and the higher the number, the greater the degree of correlation; an R-squared of .1 is basically what you get if you draw random points on a graph. Or, to put that another way:

Graph3

That’s a scatter plot of the strikeout rates and wRC+s of players from the 1961 season (i.e. the one with the “highest” correlation). Does that LOOK like a correlation to you? Hopefully, you answered no (because of the way the internet works, I can’t know what your answer was, or even if you answered); any monkey⁵ with even a basic grasp of statistics could see that those two variables aren’t connected in any way.

What, then, does this mean?

Not only are strikeouts not correlated to offensive success now, they never have been, and probably never will be. Now, can we please stop saying they are⁶?

—————————————————————————————————————————————

¹I tried looking up specific quotes, but searching “strikeout stigma” just returned some ADHD thing.

²And, of course, scatter plots reflecting such will generally be more elliptical than straight.

³In case you’re unedumacated, R-squared measures the degree of correlation between two variables. It returns a value between 0 and 1; the higher the value, the greater the correlation, and vice versa.

⁴I’m forced to say “R-squared” to avoid confusion between that and the footnotes.

⁵Really, a monkey would probably be the one drawing the scatter plot.

⁶Or were. You know what I mean.





Triple R enjoys pestering the writers at Camden Depot. He can be called R Cubed if you are so inclined. Referring to him by any other eponym may result in the infliction of great amounts of physical, mental, or emotional pain upon yourself, possibly inflicted by him, possibly inflicted by a third party. He is a terrible person. This is all you need to know about him.

9 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Jonah Pemstein
10 years ago

Strikeouts may not matter for individual stats, but they do decrease the run expectancy of an inning more than a different kind of out does (not by very much, but in the long run, it adds up). Also, players with lots of strikeouts don’t make as hard contact, which leads to both a lower BABIP and fewer balls in play, which is a bad combination. And while the r^2 may not be very high, there is still a positive correlation. So strikeouts may not matter very much, but they do matter.

evo34
10 years ago
Reply to  triple_r

Where did he say “Negative R-squared”? Also, who ever argued that strikeouts only recently lost importance? This whole article is based on a premise that has never existed.

Plagarist
10 years ago
Reply to  evo34

What I’ve heard most recently is that pundits/commentators/old people are blanching at the higher K% of lots of players, and the lack of concern on the part of those players about said high K-rates. No one has argued that it recently lost importance, but the only things I’ve seen have showed that strikeouts didn’t hurt offensive value in the past few years, so this is nice since it shows that they really never have.

Jonah Pemstein
10 years ago
Reply to  triple_r

Yeah, I meant more that even though the correlation is very small, lower k% does correlate with higher wRC+. Sorry, that was unclear.

Plagarist
10 years ago
Reply to  Jonah Pemstein

“players with lots of strikeouts don’t make as hard contact, which leads to a lower BABIP.”

Mike Napoli, with a 32.4% K rate and a .368 BABIP, would like a word with you.
And no, a strikeout does not decrease the run expectancy more than a different kind of out does. I get what you’re trying to say, but a strikeout is NOT inherently less valuable than a fly out with the bases empty, or a ground out where the runner doesn’t advance, or, or, … Yeah, in the long run, balls in play that are outs are more likely to get incidental value via a runner scoring or advancing than a strike out, but the whole point of this article is that that difference is very minor.

chris moran
10 years ago
Reply to  Jonah Pemstein

This is just wrong. Players with lots of strikeouts make better contact.

http://www.fangraphs.com/community/in-defense-of-striking-out-ideal-strikeout-rates-for-hitters/

chris moran
10 years ago
Reply to  chris moran

Referring to Japem’s post.

John C
10 years ago

There was a study in one of Bill James’ Abstracts back in the 1980s that addressed this issue, in a comment on Dave Kingman. Kingman, as everyone probably knows, struck out a lot. James was trying to confirm or refute that Kingman’s enormous strikeout totals hurt his team more than other types of outs.

He matched team stats from teams with similar offensive numbers, except for that some of the teams struck out a lot and others didn’t, and came up with two groups of teams that had something like a difference of 13,000 strikeouts, but were almost identical otherwise. To make a long story short, the impact of the 1,816 strikeouts that Kingman made, had they all been other types of outs, was likely about eight runs. In other words, a strikeout is slightly more damaging than another type of out, but by such a tiny amount as to not be worth worrying about.

And if that doesn’t convince you, then the 1991 Detroit Tigers should have convinced anyone. Last in batting average, next-to-last in hits, first in strikeouts by a huge margin–but second in the league in runs scored. They had four players who were well over 100 K’s, and three more who topped 90.