Filed under:

# How hard-hit rate can signal success at the plate

Introducing—and immediately regretting—hard-hit rate batting average.

One of the more in-vogue measures of a player’s performance lately is exit velocity. It’s relatively natural not just to marvel at the distance of a home run, but how hard it was hit. And, thanks to Statcast, we get captivating updates like these sent right to our phone:

Apart from learning that Pedro Alvarez hits baseballs very hard, we are given some extra measurements for our consideration on how hard Alvarez hits baseballs. We’re given the exit velocity, the launch angle in degrees, the hang-time—to show if it was a fly ball that carried or a line drive no doubter—and, of course, the projected distance. Instead of measuring where it hits the seats—which would be different in every stadium—you can use some #gorymath to solve where the ball would land if there were no obstructions.

More to the point though, distance of a ball is impressive, but a home run is a home run whether you hit it 450 feet or 500 feet. Unfortunately then, no matter how hard Giancarlo Stanton hits the sculpture in Marlins Park, it still only counts as one. The same can be said of launch angle, hang-time, and exit velocity.

It seems like our collective fascination of speed makes us care about exit velocity a lot more though, doesn’t it? I don’t hear about Joc Pederson’s launch angle, I hear about his exit velocity. The same can be said of pitchers. Anybody know the arm angle or release point of Aroldis Chapman off the top of their head? I bet you know that he threw 105 mph recently though.

We can’t really blame the ooh’s and ah’s of speed though. First, miles per hour is a much more universal unit of measure than degrees of launch angle, or the x, y coordinates of a release point—some people legitimately don’t speak that language. Most importantly though, it feels like there’s something to exit velocity, doesn’t it?

That’s a vague and perhaps disingenuous way to phrase it. What I really mean is the velocity by which a player swings his bat to hit a ball quickly seems not only repeatable—as a genuine skill—but as a legitimate way to measure that player’s future success as well. If Alvarez can hit one ball 115 mph, logic follows that his swing naturally precludes that type of speed and success.

That’s dangerous territory. Projecting future success in baseball—or any sport—can be fickle. But it still feels right. If you’re still unconvinced, ask yourself these questions: If Player X has an average exit velocity of 105 mph over 500 at-bats, and Player Y has an average exit velocity of 95 mph over the same sample, which player has the better batting average? Slugging percentage? wOBA? Which player will be better in the future, assuming their age and all other variables are consistent?

So, if it’s this easy in our heads for this narrative to make sense, why hasn’t someone converted exit velocity into a context that is better understood; a hard-hit batting average if you will. Fortunately, FanGraphs sorts batted balls into bins—courtesy Baseball Info Solutions—called soft-, medium-, and hard-hit. This isn’t technically exit velocity, though it does a good job at approximating a player’s batted ball ‘success.’ Until it’s easier to sort by specific mile-per-hour exit velocities, this seemed to be my best course of action.

Batting average, as far as measuring a player’s actual worth, does a pretty poor job. This shouldn’t be revelatory, you’re on Beyond the Box Score. However, one would presuppose at least a small correlation between hard-hit rate and batting average, since all it measures is a player’s ability to reach base safely once they’ve put a ball into play, while also penalizing them for striking out.

This was such a forgone conclusion to me, that I skipped this step. To be fair though, I had another motive for skipping that: batting average is not good. Not only did I not want to graph the correlation between batting average and hard-hit rate, but I didn’t want my hard-hit rate average—which would be calculated using the trendline equation—to have the extra noise of batting average in it.

Instead then, I grabbed all of the qualified hitters from 2015—141 players—with their batting average, wOBA, and hard-hit rate. I then graphed hard-hit rate to wOBA—a more complete measure of a player’s offensive worth—then did some corrections so that it looked more like batting average. Average wOBA is .320, 50 or more points higher than the mean batting average.

Good news: hard-hit rate and wOBA seem to correlate. Bad news: not that well. The relationship between 2015 major league hitters’ wOBA and their hard-hit rate is just over 47 percent explained by r-squared. That is—and I don’t mean to get too technical here—what they call ‘not great.’

That trendline gives you an equation of y = 1.094x - 0.057. Correct to batting average and you get y = 1.094x - 0.065. The top ten in the chart look like this:

Name Team AVG wOBA Hard% y=1.094x-0.065 FORM - AVG
J.D. Martinez Tigers .282 .372 .428 .403 .121
David Ortiz Red Sox .273 .379 .419 .393 .120
Matt Kemp Padres .265 .325 .416 .390 .125
Paul Goldschmidt Diamondbacks .321 .418 .414 .388 .067
Chris Davis Orioles .262 .390 .414 .388 .126
Bryce Harper Nationals .330 .461 .409 .382 .052
Mike Trout Angels .299 .415 .408 .381 .082
Miguel Cabrera Tigers .338 .413 .401 .374 .036
Brandon Belt Giants .280 .359 .398 .370 .090
Andrew McCutchen Pirates .292 .380 .394 .366 .074

That last column on the right there is how much higher the particular player’s hard-hit rate average beats his actual batting average. And it is all over the place. It’s also worth noting at this point—in case it wasn’t already apparent—that the average difference between the formula’s output of batting average and the player’s actual batting average for the sample is zero. That means the bottom ten look like this:

Name Team AVG wOBA Hard% y=1.094x-0.065 FORM - AVG
Andrelton Simmons Braves 0.265 0.290 0.229 0.186 -0.079
Didi Gregorius Yankees 0.265 0.303 0.225 0.181 -0.084
Cameron Maybin Braves 0.267 0.307 0.214 0.169 -0.098
Alexei Ramirez White Sox 0.249 0.279 0.213 0.168 -0.081
Alcides Escobar Royals 0.257 0.271 0.208 0.163 -0.094
Jose Reyes - - - 0.274 0.300 0.197 0.151 -0.123
Jean Segura Brewers 0.257 0.268 0.197 0.151 -0.106
Ben Revere - - - 0.306 0.316 0.181 0.133 -0.173
Dee Gordon Marlins 0.333 0.337 0.176 0.128 -0.205
Billy Burns Athletics 0.294 0.317 0.138 0.086 -0.208

Again, it’s all over the place and distantly deviated from the mean. This side of the chart highlights a specific set of players though: speedsters. And that makes a lot of sense. Players like Billy Burns, Dee Gordon, and Ben Revere are able to inflate their batting averages by reaching base more often on weaker hit groundballs. This is relatively well-documented for players that beat the league average BABIP as well.

It got me thinking though. At the top of the list, are players really being undervalued by batting average if their hard-hit rate indicates that they should be playing better? Or is the bottom of the list teaching us something? That under-performing batting average relative to hard-hit rate actually signals something entirely different about a player’s game?

So I re-sorted the list to grab the players whose hard-hit rate beats their batting average by the widest margins. The top five players that came up, in order, are Brandon Moss, Joc Pederson, Chris Davis, Matt Kemp, and J.D. Martinez. After a quick look through the FanGraphs’ 2015 leaderboards, four of those players happen to sit in the top ten of strikeout percentage with Kemp sitting 28th. This is the least revelatory moment of all of this to me. Batting average penalizes for strikeouts, hard-hit rate has no way of encompassing that.

As a somewhat unfortunate conclusion, we’re not left with much here. The correlation between the hard-hit rate formula for batting average and a player’s actual batting average is abysmally low—just over two percent. It would seem counter-intuitive to base a formula that combines hard-hit rate and wOBA to predict a player’s wOBA. Even further, it wouldn’t prove anything to show a correlation between x and y can lead to an okay prediction of y.

The most intriguing take aways for me then are largely about how little we know about batted ball data. While I enjoy seeing majestic homers alongside their measurements, I’m still not quite sure what to make of these measurements. In case you’re curious, the only player in which the formula exactly equalled the player’s batting average was Derek Norris. What does this mean? ¯\_(ツ)_/¯