Strikeouts: Definitely Bad, Likely Necessary

Striking out can be at once an art form as well as a temporary work of art (via Michela).

Striking out can be at once an art form as well as a temporary work of art (via Michela).

Strikeouts are bad. One thing learned so far.

More strikeouts for a baseball team result in fewer opportunities to score runs. Many good (and some bad) discussions have been held regarding strikeout rates as they have steadily increased the last few years. Major league team personnel occasionally make public statements calling for reducing team strikeouts, and with good reason for the most part. Numerous writers and fans yearn for teams to put the ball in play more often; some cite the lack of productivity inherent in striking out, while others just think balls in play are more interesting to watch. Ignoring the entertainment aspect of it, here I present some of the good and the bad associated with strikeouts, as well as the dangers in simplifying the argument to “just put the ball in play more.”

I have had a few conversations with a hitting coach friend of mine on this very topic. Back when he was a scout, the team he worked for looked at strikeout rates and how they correlated with team scoring and other offensive statistics. They found that the best teams in the league did not limit strikeouts as a key to success. The top contemporaries hit the ball with enough authority that the strikeouts did not matter, giving them more runs per hit than the light-hitting contact teams. With the change in run environment over the last decade, and the corresponding calls for playing small ball and putting the ball in play, I thought it was as good a time as any to revisit the idea and paint a new picture.

It is obvious how striking out too much can be bad for teams’ run scoring abilities. However, the potential tradeoffs in attempting to reduce strikeouts are important to remember as well. Presumably there is also a point where teams can strike out too little. If a team puts every ball in play, the quality of contact will assuredly suffer to accommodate the difficult task of making contact every at bat. Total offensive production could be diminished as a result.

While the number of singles and reaches on error may increase simply from more chances to get a ball through, so would grounding into double plays, and weakly hit fly balls and grounders. Well-hit balls would be less common as well, since hitters would have to sacrifice bat speed for accuracy. On top of that, the other teams would undoubtedly recognize the tendency and adjust. Pitchers would be able to take advantage and further exacerbate the loss of efficiency.

For some data, let’s first explore team strikeout rates and how they relate to other aspects of offensive production. Pooling together all the team seasons from 2002 to 2013, we can look at the simple correlation between strikeouts and runs scored.

02-13 Team Stats
Not a strong relationship, but recent history has shown that the highest scoring teams have, on average, struck out less often than lower scoring teams. Higher strikeout rates are associated with lower on-base percentages (OBP), slugging percentages (SLG), and by logical extension, weighted on-base averages (wOBA). There are simply fewer chances to accrue positive outcomes when strikeouts are more prevalent. However, if a team is able to do a large amount of run-scoring damage with those chances, it still can have a productive offense despite high strikeout rates.

Increased team strikeouts are also correlated with several positive offensive developments. Grounding into a double play happens at a lower rate when more at bats end in a strikeout, and more fly balls turn into homers for teams with higher strikeouts. This is because teams are more accepting of high strikeout hitters as long as they hit the ball hard enough to offset the missed opportunities.

Team K rate & offensive statistic regression analysis
Stat Correlation (R)
OBP -0.45
SLG -0.29
wOBA -0.36
GIDP -0.47
HR/FB 0.19

Harder, more productive contact could be related to higher strikeout rates. Even if this were a stronger relationship, this is certainly not a pure causational effect, since players who are able to reduce their times grounding into double plays and hitting infield fly balls  while keeping a decent home run rate will be given a longer leash regarding their strikeouts. However, it is certain that there is a tradeoff with strikeouts that goes both ways. Striking out too much without a corresponding increase in run scoring on balls in play is obviously bad, but putting every ball in play at the expense of hard contact is also undesirable.

This data set does not show a relationship between strikeout rate and trajectory off the bat or finding holes in the defense. Team strikeout rate has next to no correlation (R-squared < 0.01) with batted ball types, batting average on balls in play (BABIP), isolated slugging (ISO), walks (BB), infield fly ball percentage or infield hits (IFH). I was especially surprised that team ISO did not have a stronger correlation here, if high strikeout rates are indeed a natural complement to hard contact. Just for comparison’s sake, here’s what the same graph from above looks like for the 1992-2001 period:

In the higher run-scoring environment of the late 1990s and early 2000s, strikeout rate indeed proved to be much less important in regards to scoring runs. This certainly lends some credence to putting the ball in play more given that there are now fewer high-octane offenses. Before we drink the Kool-Aid, let’s dig further into the individual player numbers for more information. These numbers include all of the individual qualifying hitter seasons over the same stretch of time as above (2002-2013):

Individual K rate & offensive statistic regression analysis
Stat Correlation (R)
OBP -0.07
SLG 0.19
ISO 0.40
wOBA 0.10
GIDP -0.22
IFH -0.27
HR/FB 0.53
IFFB% -0.16
GB% -0.28
FB% 0.31
BABIP 0.08
BB% 0.28

The strongest relationships are to the power numbers: home runs per fly ball, isolated slugging percentage and fly ball percentage. Walk percentage shows up with a slight positive correlation to strikeouts as well, due to going deeper into the count to end an at-bat in either fashion. Grounding into double plays and infield hits both go down with increasing strikeouts, just as above. It does look as if striking out more is a natural complement to hitting the ball hard and far by these numbers, even more so than in the team numbers.

The only problem here is the selection bias inherent in looking at a group of major league regulars. Hitters who hit the ball harder or farther than average are not taken out of the lineup as quickly if their strikeout totals are high. The vast majority of high strikeout hitters may be forced out of the lineup before they are able to gather a full season of at-bats. Power hitters are valuable in part due to their increasing rarity relative to the number of contact hitters, allowing more of them to accrue plate appearances even if they have other deficiencies (e.g. strikeouts). It is also not necessarily true that hitting for power requires strikeout numbers to rise.

There is an argument to be made that strikeouts are simply a necessary evil that accompanies hitting a baseball with authority; another is that teams are more tolerant of strikeouts if they also happen to come with hard hits. I’m comfortable assuming both conclusions to be somewhat true. This leaves us with some gray area when it comes to the requests expressed by many fans and baseball people. On the team level, strikeouts clearly can diminish run-scoring capabilities (at least in the recent run environment). Since the individual level is the only realistic place to make changes with existing personnel, let’s analyze what processes lead to individual strikeout rates.

Using FanGraphs’ handy collection of plate discipline numbers, we can see how players can improve their strikeout rates. With this, we are looking for hitters who perhaps need to lay off pitches outside of the zone, or swing more at in-the-zone pitches, or simply make more contact. For our individual hitter sample covering all the years for which we have plate discipline data, here are most of the available metrics plotted against strikeout rate:

K v Plate Discipline
When it comes to what is traditionally known as plate discipline—swinging at strikes, not swinging at balls—there really is no correlation to strikeouts in the aggregate. This is an important point, since many player development programs preach to their hitters about knowing the strike zone and hitting “good” pitches (i.e. strikes). The problem with that is not every hitter has the same happy zone in which he consistently hits balls hard. I don’t think anyone complains when Miguel Cabrera swings at a pitch inside off the plate.

Surely some hitters could improve based on their own unique strengths and weaknesses, but there is not enough evidence here to prove hitters could decrease strikeouts by being more disciplined at the plate based on the rulebook strike zone. The only useful pieces we are left with here are swinging strike percentage and contact percentage. Both describe the same skill: making contact and putting the ball in play.

This brings us back to the original problem: Simply putting the ball in play more has many potential side effects. For a hitter with poor contact skills, it is not a no-brainer decision for him to sacrifice well-hit balls for fewer strikeouts and more balls in fair territory. Even that is a pretty big hypothetical, since we have to presume hitters can choose to put the ball in play more. Major league pitchers throw a lot of crazy stuff, and even bunting can be difficult for some great athletes.

Though we do not have data documenting hitters’ thought processes, we can look at a situation where hitters tend to try putting the ball in play more often. In general, with men on base, hitters have an incentive to move runners along in an attempt to score runs. This is partly due to runs batted in still being a popular stat among on-field crowds, as well as the systemic appreciation for giving up an at-bat to move runners closer to home plate, among other reasons. By looking at the results hitters achieve with men on base contrasted with no one on, we can get an idea of the limits a change in approach can have on strikeouts and how much benefit or penalty a change can cause.

I gathered all the hitters from just the 2013 season who had at least 50 plate appearances with runners on and 50 with the bases empty. I looked at the change in strikeout rates between the two scenarios for each hitter (for a sample size of 435 hitters) and the corresponding differences in offensive production across multiple categories. Over the whole sample, the average change in strikeout percentage was a 1.49 percent drop with men on base (s = 4.90 percent). Of the 200 hitters with the most plate appearances, the high and low values were a 6.40 percent increase and a -10.10 percent decrease, for what it’s worth. Here are a few of the stats from the regression analysis:

2013 Individual hitter change in K rate Vs. change in offensive statistics between bases empty & men on base situations
Stat Correlation (R)
OBP -0.12
SLG -0.12
ISO -0.07
wOBA -0.13
HR/FB -0.02
BABIP -0.03

Nothing pops up here. I tried to limit it to players with a higher number of plate appearances to eliminate some of the noise from the lower totals, but the correlations were the same or worse. Batted ball profile changes were also of little significance in regard to strikeout rate change. This is a dead end, at least for my brain.

Perhaps this article offers some leads that others can explore to better explain these relationships, since I think the general understanding of strikeouts in baseball is still lacking. With a likely tradeoff existing between swinging strikes and quality contact, asking hitters to make more contact is not a viable solution for teams looking to improve their whiff-prone offenses. Improving pitch selection as it pertains to strikes and balls is also not supported by this data as a useful change.

When it comes to strikeout rates, I feel like what you see is what you get. Pitch tracking, eye-hand-body coordination, pattern recognition, concentration, stress management, and reaction time are all much more important and intangible skills than swing rates that shape hitter contact consistency. For teams looking to evaluate and improve contact rates, these are just a few of the many avenues to be investigated to employ more elite hitters. Until then, either leave the strikeout guys alone or trade them to someone else who will appreciate the positives.

References & Resources

  • All statistical data pulled from

Dan is Fangraphs Lead Prospect Analyst, living in New York City. He played baseball for four years at Franklin & Marshall College before attending medical school. Follow him on Facebook and Twitter @DWFarnsworth.
Newest Most Voted
Inline Feedbacks
View all comments
Peter Jensen
10 years ago

You may want to rerun your analysis with the teams separated by league.

10 years ago

I don’t know, Cain, Lincecum, and Bumgarner do much better when they are striking out more. :^)

Some points. You point out Walk Rate as a slight positive correlation at 0.28 while simultaneously noting Flyball Rate as one of the strongest, which was 0.31; I understand your point, but it just struck me wrong. Personally, I would have pointed out the two as the strongest correlation, then note Flyball rate and Walk rate as relatively high (could have thrown in SLG too), and then make the point that all three has been associated with power hitters (three true outcomes).

I’m wondering if you are muddying the waters by throwing in all hitters. You want to see what the relationship between strikeout rate and all those offensive stats, and found a number of relationships. Well, shouldn’t you be removing all hitters who are good at avoiding strikeouts then? The contact hitters? You can’t see what the value of striking out is if you throwing in the data of hitters where the hitters purposefully avoid strikeouts.

So you might want to split the study set into three groups: high, middle, and low. Then throw out the middle and do your analysis on the high and low, and see what you find. Or even go to deciles and see how each group does in your analysis. Not sure what will come out, but that’s what I would suggest.

About your analysis looking at hitters with runners on and bases empty, that’s the traditional way analysis was done in baseball to see if there was any clutch hitting in baseball, and the results convinced many that it does not (it does). So I’m not surprised that there is very little correlation. And that’s why it is a dead-end.

I don’t think it does hitters justice to say that because they like RBI’s, they have an incentive to move runners along. They like RBI’s because it means that their team now has one (or more) runs than they had before. The idea is ideally to get a hit and drive in the run, but if not, at least get the runner up another 90 feet.

And not every team believes that giving up outs is the right thing to do. The notion that sacrificing is the wrong move has been around at least as long as Earl Weaver was a manager, he famously hated giving up outs via sacrifice bunts or getting caught stealing.

I think the overarching “aha” of your research is that strikeouts are indeed bad for run-scoring, the more of it there is, the less teams are correlated to score. The way strikeouts are mitigated is when the batter compensates by getting a lot more extra-base hits and/or by walking (like the three true outcomes). Though I would think that most of us already knew all that.

Lastly, here’s a point I would make: the best pitchers are known, via sabermetric studies, to be those who strikeout more batters and who have a good to great K/BB ratio. So clearly, strikeouts are bad. If it weren’t, then this wouldn’t be so, I would think. Then it becomes a matter, for the hitter, to provide value that makes up for that, in his other PA’s. So they aren’t necessarily necessary, just more taking the bad with the very good.

10 years ago

>> “If a team puts every ball in play, the quality of contact will assuredly suffer to accommodate the difficult task of making contact every at bat.”

Can that be tested with simulated games? Suppose you have two identical teams playing, except for one team every swinging third strike is replaced by a randomly selected ball in play that is hit with poor contact. (I might exempt power hitters who get a lot of home runs because I don’t want those players to change their swing and their strikeout rate should not change in the simulation. ) In other words, for the player getting the simulated swinging third strike, look up his record on balls in play that are launched from the bat at relatively high or low angle, assmuming that shows his ability when he makes poor contact, and replace his strikeout with one of those other outcomes. Does the team get better when this is simulated? I know this is flawed, and I intentionally left some things vague, but at least it sounds like something fun to play around with.

10 years ago

The graphs for 1992-2001 and 2002-2013 are the exact same image with different titles.

Matt G
10 years ago

Good article, but for god’s sake please label your axes.

Paul G.
10 years ago

Another way to attack this issue is focusing on player speed. Fast players should benefit more from balls in play as they can turn weak ground balls into infield hits more often. So perhaps focusing on contact would not work especially well for an Adam Dunn or Jose Molina, but it might do wonders for Austin Jackson or someone like him.

10 years ago

Don’t be ridiculous. Just have every player strikeout less and hit the ball harder and further. Simple solution.

Peter 2
9 years ago

It has been intuitively understood that a team can accept a player striking out a lot—even a ton—if and only if that lack of contact trades off with huge power numbers. That’s why Ryan Howard, Adam Dunn, and Mark Reynolds were viable offensive players in the 2000s, and why you don’t see a lot of players who strike out 200 times and hit 2 HRs. Notice, however, that the margin of error for these sorts of high power/low contact players can be quite slim, and the decline can be steep at that tipping point. When Dunn’s power slipped even a little bit, even 34 HRs last year wasn’t enough to bump him past replacement level. And think of how quickly Richie Sexson fell off the face of the earth.

John C
9 years ago

What I get out of this is if teams want to improve their offense, get better hitters. The best hitters command the K zone, hit for power, get on base and don’t strike out too much…e.g. Miguel Cabrera and Mike Trout. I’m not sure you can teach mortal players to hit like those guys, but there is a band of all around great hitters who offer a high percentage of Cabrera/Trout performance…perhaps Joey Votto is a good example.
Great hitters aren’t cheap, so after that teams need to pick desirable component parts to assemble their lineup. High ISO/High BB/High K players, High ISO/Low OBA/High K players, High BABIP/low K players, High OBA/High K, etc. It kind of reminds me of the most recent Billy Beane approach to lineup construction.
Another thought before I go…I wonder if line drive% should be included in this somehow? I haven’t given it much thought, yet.