March 17, 2015
Counts Are Crushing Offense
At-bats are baseball’s fundamental building blocks. Good at-bats make for good games. Ninety percent of the suspense and rising action in baseball lives in the episodic unfolding of a plate appearance. Of course, not all plate appearances are created equal. In fact, it’s a pretty radical time in MLB history, in terms of the anatomy of a trip to the plate. On Friday, I wrote a piece that explored the surprising way the Twins attack the first pitch—or rather, don’t. Today, I want to dig more into at-bats, from beginning to end, to see how their inner workings are different than they were 25 years ago—or 20, or 10, or even five.
First, let’s revisit what happens right away. The first pitch is a high-leverage offering. In fact, what happens on that pitch has never mattered more. In 2014, batters hit .336/.341/.526 on the first pitch. If you’re unfamiliar with how these things work, maybe that sounds terrific. It’s not. Of the 27 seasons for which we have data, that .867 OPS ranks 16th. Compare it to the overall league run environment, though, and it positively shines.
That shouldn’t surprise us. One reason for the drop in run scoring over the past half-decade is the soaring strikeout rate, and you can’t strike out on the first pitch. When you’re watching a game, know that the first pitch is an opportunity for the offense.
It’s an opportunity they can ill afford to miss, too. Forget production relative to the league norm. Absolute production is endangered (to the tune of a sub-.600 OPS) as soon as the pitcher gets ahead 0-1.
Of course, putting the ball in play on the first pitch (a good thing) requires swinging, which puts one at risk of falling behind (a bad thing). What if one takes the first pitch, as the Twins do so often? That same piece from Friday showed that pitchers are throwing so many first strikes that taking is a worse strategy than ever. Only 39.3 percent of all plate appearances went to a 1-0 count in 2014, the second-lowest rate on record. And when they did get there, players only had a .792 OPS the rest of the at-bat, a 20-year low.
Now, the shapes of the two graphs above aren’t all that different. The relationship between a 1-0 count and an 0-1 count is not much different today than it was 10 or 15 years ago. Both expected OPS figures are just lower, because after that first pitch, more at-bats overally proceed in the wrong direction for a hitter than they did in the past. Relative to overall production, falling behind 0-1 isn’t a more costly problem than it used to be.
Part of that, however, is the fact that falling behind 0-1 is more common than it used to be. Another part is that, once 0-1 happens, batters are getting more defensive and swing-happy, and that’s leading to more 0-2 holes out of which to work.
Swing Rates and Frequency of Reaching 0-2 After 0-1, 2008-14
The league’s OPS was right around .700 last year. Hit the first pitch in play, and you got a bump nearly to .900. Get ahead 1-0, and you had a bit less than a .100 bump from when you put on the helmet. Fall behind, and you took a little more than a .100 hit.
In only about every ninth plate appearance was the first pitch the deciding one. So what happens on the second pitch? As you might imagine, it depends on how the first pitch went.
Batting Performance By Count After Two Pitches, 2014
Percentage of PA Reaching 2-0
When I say that that first strike is a tough hole out of which to dig, I mean it. After even reaching a 1-1 count, an average batter in 2014 could expect to strike out a quarter of the time. As recently as 2005, that number was more like 21 percent, and it was a whole lot more livable.
This is why we see more hitters swinging away down 0-1. The ever-growing strike zone puts them in considerable danger of falling down 0-2 even if they take that pitch, but more importantly, they’re still fighting their way out of a corner if the count does even out. There’s too little to gain by trying to get to 1-1, and too slim chances of getting to any more favorable count to discourage those swings. Unless the strike zone really does get smaller in 2015, batters should continue to take an historically aggressive approach to 0-1 pitches.
I mentioned that only one in nine plate appearances is over after the first pitch, one of the lower numbers on record. The percentage taken care of in a 1-0 count is even lower, both in absolute terms (6.8 percent of all PA end with that count) and as a percentage of all at-bats (that 6.8 number is the lowest ever). Still, plate appearances lasting two pitches or fewer counted for 27.2 percent of all plate appearances in 2014. The most common number of pitches for any given plate appearance to last is two. From here, an at-bat can splinter in a thousand directions. The prudent thing to do, rather than wade through all of those possibilities, is to widen our lens. Let’s ask this simple question: How common was it, in 2014, for a plate appearance to end with the batter ahead in the count?
The answer is 33.0 percent, and for what it’s worth, that’s the smallest share of plate appearances that have ever ended under those circumstances.
More at-bats are ending with the count in the pitcher’s favor, which is having precisely the effect you would think on league-wide offense.
Here are the overall league batting lines for 1988 (the first year for which we have data on the count), 2014, and 2001 (halfway between the two), along with the lines for only those plate appearances that ended with the batter ahead:
I think this kind of urgency, this sense that the batter is racing the clock from the moment he enters the batter’s box, makes for bad at-bats. The only good solution, though, is for some fundamental change to come to the game, something that will restore a hitter’s ability to assert control of the count at times. For my money, that ought to be automated umpiring, or at least, a codified shrinking of the strike zone. I have no doubt that today’s pitchers have the best control of any set of pitchers in baseball history, and I get that batters swung at a higher rate in 2014 than they had since 1990. Still, the league’s aggregate walk rate last season was the lowest since 1968, and second-lowest since the Dead Ball Era. When 1968 comes up, something has gone wrong, and needs to be fixed.
In the meantime, there are teams and individuals taking unique approaches to this problem, and we should continue to watch them, to see if any bear any particular fruit. The Astros (extremely aggressive early in counts, extremely passive late in counts) seem to be buying into the trend arrows. They want their hitters stepping into the box ready to rip the first good pitch they see, and if one doesn’t come early, they want their guys waiting hurlers out, hoping those pitchers will let them get back into the at-bat. Other teams are more inclined to try to get their batters ahead first, and then cut them loose.
Ultimately, most differences in expected outcomes in baseball are very, very small. We talk about lineup construction and bullpen optimization until we’re blue in the face, but the expected changes in runs scored or games won based on the changes going on in those arenas are minute. What we fail to see, sometimes, is that the smallest waves of change can have the biggest consequences. The leverage of the count is just one example, but it might be the best one. Each pitch (especially early in the count) wildly swings what we ought to expect out of an at-bat, but we’re too busy viewing the at-bat as the single event between the one before it and the one after to take notice.