There's gotta be a reason for the strikeout epidemic ... right?
At the beginning of the 1980s – a time which my wife constantly reminds me is now preserved on the oldies station – the average baseball game featured 9.6 strikeouts. Total. For both teams. Indeed, from 2008 – a time which my wife reminds me is now preserved on NOW! That’s What I Call Music! #835 – until today, the average MLB game has added an additional two strikeouts. We’re in the middle of the greatest strikeout era in professional baseball history, with teams now combining for in excess of 15 Ks per game so far in 2014.
What happened? How did we get this way? Can’t anyone hit the ball anymore? I propose that we test a couple of theories and see whether they hold up in light of the evidence.
Warning! Gory Mathematical Details Ahead!
Theory #1: Pitchers these days! The new ones throw nothing but heat and it’s impossible to hit the ball nowadays.
I looked at all plate appearances from 1993 to 2013. Starting in 1993, I looked for all pitchers who pitched their final frame that year, and figured out what percentage of hitters they struck out. Then, in 1994, I looked for all pitchers who made their debut, and again, found the percentage of hitters whom they faced who did the walk of shame back to the dugout. I took the difference between those two numbers, and the did the same calculation for 1994 into 1995, 1995 into 1996, etc. I limited the sample to plate appearances against batters who were active in both years, to make sure that it wasn’t a bunch of new hitters with holes in their swings that’s really to blame for the difference.
We’d actually expect the difference between those two numbers to be positive, because teams should be getting rid of the duds who don’t strike anyone out and replacing them with guys who can record a K. But, over time, If pitchers are getting just so much better, we’d expect to see that difference get bigger. Here’s a graph of that difference over the past 20 years.
What’s striking here is that while debutants are generally more likely to strike hitters out than the retirees whom they replaced, in recent times, the difference between these two has not been especially large. I don’t think that we can fully pin the blame for the recent surge in strikeouts on the pitchers, so…
Theory #2: Hitters these days! The new ones do nothing but strike out, unlike ye olde generation which at least knew how to put the ball in play.
I basically ran the reverse of what I did above, looking at retiring hitters from one year and then debuting hitters in the next, when they faced off against pitchers who pitched in both years. I suppose that we should expect this to be negative, as teams should, in theory, be replacing hitters who tend to strike out with those who don’t. But instead we see this:
It seems that as late as 10 years ago, teams generally replaced retiring hitters with hitters who struck out just as much or slightly less often. But since those days, there’s been a steady upward slope. The hitters who retire are generally less likely to strike out than those who replace them. It seems that at least some of the growth in strikeouts has come from the introduction of hitters who strike out more.
That doesn’t seem to make much sense. Why on earth would hitters (and the teams that employ them) do something that increases the rate at which they strike out? It makes sense from a pitching standpoint, but… hitters?
Theory #3: Hitters don’t have contact skills any more. It’s all just “swing real hard in case you hit it.”
I’ve looked at this theory before in a previous article. What’s interesting to note is that swing rates have stayed pretty consistent over the years Contact rates have fluctuated up and down, and lately have been down, but the spread hasn’t been giant. The highest and lowest seasonal rates are within a few percentage points of each other.
And if hitters are selling out for power, then they aren’t doing a very good job. I looked at a stat called SLGCon (slugging percentage on contact), which looks at slugging percentage in plate appearances where the ball was put into play. It’s a crude measure of the power that hitters generate when they hit the ball. That’s gone up and down over the years, and hasn’t tracked the rise of strikeouts.
If hitters aren’t swinging appreciably more, aren’t swinging and missing appreciably more, and aren’t selling out for power (or at least succeeding if they really are trying), then what on earth could possibly be driving the strikeout scourge?
Theory #4: The silent killer: strike one.
Let’s set up a couple of other developments that have happened in the past few years. Plate appearances are lasting longer (as measured by average pitches per PA). That’s unquestionably true. In the last 20 years, hitters have seen about 0.2 pitches more per PA. To get through the lineup three times, a pitcher now needs to throw about 5 more pitches than he did in 1993.
This matches up well with the general wisdom that teams are now trying to extend at-bats in order to drive up the starter’s pitch count. Since most starters enter the game with a pitch limit stamped on their forehead, it makes sense to try to get him through his allotted pitches and into the soft underbelly of the other team’s bullpen. If you knock the starter out in the fifth inning, the other team will need to use a couple of (less good) relievers whom they would not otherwise have used if the starter made it to the seventh. But if hitters aren’t going to swing as much, why not fill up the strike zone?
Again, from 1993-2013, I looked at the first pitch in each plate appearance and looked at the rate at which players swing in each of those seasons. There’s been a fairly steady decline over the years (although in fairness, a tiny rebound over the past few seasons.)
Interestingly enough, the graph for 1-0 counts (which I have not shown) looks similar with swing rates falling, although not as much. On 0-1 counts (again, not shown), swing rates have actually gone up. It seems that hitters have adopted a philosophy that says “I’m not swinging until you throw me a strike.” Pitchers have seemingly responded with a resounding “OK, if you insist.”
Going back to 0-0 pitches, this time only looking at the ones where the batter did not swing, I looked at the percentage of pitches which ended up as strikes vs. ended up as balls. In an era where hitters are taking more first pitches, pitchers are responding with the obvious. If he’s going to let me have strike one, I’m going to take it.
As a result, we’ve seen an increase in the percentage of plate appearances which go to 0-1 after the first pitch, rather than 1-0. (Plate appearances in which the ball was put into play on the first pitch are excluded.)
Now, what’s the difference between a 0-1 count and a 1-0 count. It turns out to be a lot. This is a graph of the chances that a plate appearance will end in a strikeout, based on whether it begins with a 0-1 count after the first pitch versus a 1-0 count. We see that the lines run roughly parallel to each other, and that in all years, a plate appearance starting with 0-1 ends in a strikeout more than 10 percentage points more often than one that starts 1-0.
Now, there may be bias in that players who would be striking out more often anyway are more likely to see a 0-1 count on the board. But, it can’t help that hitters are hitting from behind in the count more often. A smaller, but similar effect appears when looking at plate appearances that start at 1-0, and then branch out to either 2-0 or 1-1. On 1-0, swing rates are down, called strike rates are up, and it’s easier to strike someone out after a 1-1 count than a 2-0 count. All of this does not explain the entire increase in strikeout rates, but it is most certainly contributing. It might be that the emphasis on driving up pitch counts (and defining “discipline” as taking a lot of walks) has backfired a bit and led to more strikeouts.
Theory #5: Pitchers have better put-away stuff now. Once the batter gets to two strikes, he’s done for.
There is mixed evidence that pitchers are getting better at putting hitters away. I looked at outcomes from two strike counts (0-2, 1-2, 2-2) and the first pitch that was part of that count. (A foul ball, of course, means a do-over, but we aren’t looking at take two.) I present the results for 0-2 counts because they are representative.
Hitter are swinging more on 0-2 counts since 1993, and especially since 2000. (This pattern repeats in 1-2 and 2-2 counts).
When they don’t swing though, hitters are not taking more called strikes. In fact, the rate of called strikes is going down, in general, in two-strike situations. If pitchers are trying to freeze hitters, they aren’t doing so well.
Pitchers are, on the other hand, having better luck with getting swings and misses with two strikes on two strike counts, and the effect size has been bigger than the overall effect for swings and misses. Then again, hitters who end up in more 0-2 counts are more likely to have more swings and misses to begin with, so our sample might be biased.
Maybe what’s more important is this graph, the percentage of plate appearances which have ended up in a two strike count.
All the put-away stuff in the world won’t help you strike a batter out if the plate appearance doesn’t get to two strikes.
Can the Answer Be That Easy?
It’s tempting to look for the answer to the strikeout scourge in terms of what pitchers are doing. Over the last six years, we’ve seen the introduction of the Pitch F/X system which catalogs the speed, spin, movement and location of just about every pitch thrown in a MLB game. We have such rich data. Surely the answer is in there, right?
Looking at things in an historic context, we start to see that there are probably a couple of factors at work driving up the strikeout rate. In the name of driving up pitch counts (and perhaps the equating of “discipline” with taking a lot of pitches), hitters have become more passive and as such, pitchers are taking advantage. (Why wouldn’t they?) This leads to an upswing in hitters having to hit from behind in the count and there is a world of difference between hitting ahead and hitting behind. It doesn’t seem that hitters are gaining any particular advantage in terms of power from this new approach, and a simple look at the number of runs scored per game will show that the strategy hasn’t produced such great results. Pitchers do seem to have been improving in their abilities to get swings and misses (or is it that hitters are getting worse at this particular talent?). The fact that the strikeout rate seems to be more closely related to the entry of new batters into the game suggests that maybe the batters share more of the responsibility for rising strikeout rates.
And yes, maybe the idea of driving up the pitch count as a strategy has backfired. While the aesthetics of the strikeout are debatable, the reality is that we’ve also entered a new era of lower scoring games. Since the job of the offense is to score runs, maybe teams might want to re-examine whether their plans have spawned unintended consequences … like an epidemic of strikeouts.