The value of a batted ball

Here at The Hardball Times, we like to talk about batted ball statistics. We think they can be useful and illuminating and there’s still a lot to that can be learned about player performance by looking at batted ball numbers.

But the question that always remains is: how valuable are the numbers we’re presenting? They might be interesting, but do they mean anything? How much stock should we be putting into them?

For example, a few weeks ago, Dave Studeman showed that Manny Ramirez is getting a lot less out of his outfield fly balls than he has in previous years. But does that mean that Manny isn’t hitting the ball as hard? Or has he simply been unlucky? Of course, the answer to that question will give us vastly expectations about Manny’s future performance.

So today, I’ve decided to look at whether or not players have any control over the value of their batted balls, and in addition, we’ll check into whether there is a correlation between how often a player hits a type of batted ball and how much value (as defined by linear weights) he gets out of it.

Line drives

Line drives are definitely the hardest type of batted ball to understand. On the one hand, they are the most valuable, as they turn into hits three-quarters of the time. On the other, they are extraordinarily unstable—a hitter who leads the league in line drives one year could be among the trailers the next. That can’t be said about many statistics.

Now generally, we would expect players to get more of what they’re good at—so fast players will hit a lot of ground balls, guys with a lot of power will try to put the ball in the air, etc.

But at the same time, you would think that everyone does well when they hit line drives, so that rule might not hold true here. After all, if Frank Thomas starts hitting a lot of grounders, he will suck, but can you name a player who would do badly if he started hitting liners all the time? Exactly.

And so as you might expect, there is no correlation between the value of a player’s average line drive and his overall line drive rate. In other words, the outcome of a player’s liners is random. Or so it seems…

Here’s where it gets real weird. If we take all players with at least 350 plate appearances in each of two consecutive seasons between 2003 and 2006, we find that the correlation between the value of a player’s line drives in one year and the next is 0.25—not great, but very much significant.

What can explain it? I doubt it’s that some players get more out of their line drives than others. That would certainly be a fascinating result, but it doesn’t make much sense to me. If you can explain why that may be, please e-mail your answer.

The only explanation I can think of is park effects. We know that line drives are more valuable at Coors Field than they are at any other stadium. So it’s possible that park effects are responsible for this correlation.

I guess I’ll have to calculate batted ball park effects for a future column and find out if my explanation holds water, but until then, it’s important to keep in mind that if a player is getting more out of his line drives than he has in previous seasons, there may be something to that.

Outfield fly balls

Outfield flies are a bit different from other batted ball types because they include one very stable event: home runs. Whereas what happens when a ball is put into play can be subject to a lot of luck, home runs are very consistent from year-to-year. You won’t see Juan Pierre jacking 40 balls out of the park any time soon.

But there’s more to it than just that: If you take out the home runs, outfield flies are actually a very bad play. Yes, you might get a double, but usually, outfield fly balls in play just turn into lazy outs. So for a fly ball hitter, it’s really one or the other: Either you hit a lot of home runs or you make a lot of outs.

Unsurprisingly, then, the number of outfield fly balls a player hits is pretty strongly correlated with the average value of each of his outfield fly balls—to be exact, the “r” is 0.41.

And again, without much surprise, the year-to-year correlation of the value a player derives from his average outfield fly is also very high, 0.66. In other words, there is likely something to worry about with Manny. Players are very consistent in how much they get out of their outfield flies and when they decline precipitously, that’s probably at least partly real. Manny is still a great hitter, but it’s unlikely that he’s the same hitter he has been throughout his career.

A Hardball Times Update
Goodbye for now.
Ground balls

We would expect to see the same effect here as we do for fly balls. After all, ground balls only work out well for those hitters who can regularly run them out. And speed is a very stable variable.

But the results tell a very different story. The correlation between a player’s groundball rate and the average value of his grounders is just 0.16—statistically significant, but not a very high number. Moreover, the correlation between the average value of a player’s ground balls in one year and the next is 0.31, not much higher than the same relationships for line drives.

So how could that be? Is extreme groundball hitter Ichiro just really lucky? Of course not. These correlations are a lesson in statistical noise. As I wrote earlier, a lot of funky stuff can happen when a ball is put into play, and all that random variation will drive the year-to-year correlations down.

Let’s also remember that Ichiro’s batting averages have proven to be quite variable, swinging as much as 60 points from one year to the next. Yes, Ichiro can consistently run out grounders (he’s among the league leaders in infield hits every year), but that only matters if his ground balls are hard to field.

A foot can be the difference between a hit and a routine out. Obviously, sometimes the ball will go right at the shortstop and sometimes, the shortstop will have to move just enough for Ichiro to leg it out. So of course, there will be a lot of randomness inherent in the data.

Still, we should take note of these correlations because they are important. If you see that a hitter just isn’t getting as much out ground balls as he has in the past, that doesn’t necessarily mean that he’s slowed down. He could just be getting unlucky.


Bunts, of course, are the most discretionary batted ball of all. You’re not going to bunt unless you want to. So of course, you would expect an almost perfect correlation between bunt rate and the value a player gets out of his bunts. You would be wrong. In fact, the correlation is -0.01 and completely insignificant.

Okay, that sounds wrong and I think it is. My guess as to why is that there are a lot of players who barely ever bunt providing a lot of statistical noise. So what happens if we only look at players who bunt in at least 5% of their balls in play?

Well, the correlation rises, but only 0.22, which is still lower than I would expect. My best guess is that there is still a lot of random variation in these numbers because bunts are so rare, but this correlation is definitely something to keep mind when you’re asking if the players bunting are the ones who should be.

The year-to-year correlation of the run value of bunts is also pretty low, 0.19. It rises to 0.35 if we restrict ourselves to players who bunted in at least 5% of their balls in play in both years, but the sample size also drops to a tenuously low 22. The low correlation does provide some support for the random variation theory, since running out a bunt is a pretty obvious skill.

Concluding thoughts

So what have we learned today? I’m not totally sure that we have learned anything, except perhaps the importance of randomness in statistics, which of course is an important lesson in it of itself.

However, I do think we can make some tentative conclusions about the value of looking at the value hitters derive from their various batted balls. Clearly, a lot of what we’re seeing when we look at those run values is simply randomness and should be ignored. But hitters definitely have some control over the outcome of their batted balls and clearly, they generally stick to what they are best at.

Comments are closed.