A More In-Depth Look At Curveballs: What Matters?

I initially began exploring this subject here.

In my last post, I looked at which pitch-tracking measurements– velocity, spin rate, horizontal and vertical movement– most influence the effectiveness of a pitch. I chose to look exclusively at curveballs. This post will do the same. Soon, I hope to explore other pitch types.

The question I kept asking myself while writing the first piece was: How do I measure the effectiveness of a pitch? I chose BP’s True Average. Others have looked at RE24 (the change in run expectancy following a pitch), and home runs allowed and swinging strikes.

Not too long ago, I was down at the Winter Meetings interviewing for some jobs. I had brought some of my research down, including the piece linked above, in attempts to land one of those jobs. In short, I asked some big-league employees what they would use to measure the effectiveness of a pitch in a study like mine. They said: swinging strike percentage, strikeout percentage, and ground ball percentage.

So yes, that’s what I’m going to do here. Similar to last time, the independent variables are velocity, spin rate, horizontal movement, and vertical movement. Not similar to last time, I will look at pitchers who have faced at least 150 batters, and have thrown at least 200 curveballs, since the beginning of 2014. I’ve also included knuckle-curves in the sample as well.

First up is swinging-strike percentage which correlates pretty well with strikeouts and is a good indicator of raw stuff.

This shows that velocity, or start_speed, is the only measurement that affects swinging-strike percentage. Spin rate and vertical and horizontal movement were insignificant. An increase of 1 mph increases swinging-strike percentage by .115 percentage points, on average. Now, let’s make this relationship a bit more visually appealing.

Interestingly, you can see that velocity has little influence on swinging-strike percentage for pitchers who throw under roughly 79 mph. For those who throw above 79, you can see a pretty clear relationship– the harder your curve, the more likely the batter will swing and miss. Note- the coefficient from above (.115) is an average from all velocities. But you can see rather clearly from the plot that the positive correlation exists only when the velocity is at least 79. The coefficient after running the same regression with start_speed >= 79 is .656. In other words, for pitchers whose average curveball velocity is at least 79, a 1 mph increase in velocity increases swinging-strike percentage by nearly 2/3 a point.

Next up is strikeout percentage.

Not surprisingly, what was true for swinging-strike percentage is also true for strikeout percentage. A 1 mph increase in velocity increases strikeout percentage by .385 percentage points. The same 79ish mph cut-off point exists here too.

Lastly, we have ground ball percentage. While pitchers have a tough time controlling whether a batted-ball is fielded or drops for a hit, they have proven to be able to control the direction of the ball once it leaves the bat. Some pitchers are of the ground-ball type and some of the fly-ball type. Pitchers who induce more ground balls tend to be more effective because, well, you can’t hit a home run on a ground ball.

Well this isn’t as exciting as I would have hoped. Velocity is the only significant factor here too. An additional 1 mph in velocity increases ground ball rate by nearly half a percentage point (.489). What is interesting though, is the visual relationship.

Unlike swinging-strike and strikeout percentage, which velocity only affected after 79 mph, a positive relationship exists between ground ball percentage and velocity right from the get-go. You then see a peak ground ball rate of roughly 45 percent at just over 80 mph before leveling off.

So, the harder your curveball the better if you’re seeking strikeouts and swings and misses. And if you’re seeking ground balls, there seems to be a sweet-spot at just over 80 mph.

To recap:

• Pitch Type: Curveball
  • Swinging-Strike%
    • What matters? Velocity, only greater than 79 mph. The faster the better.
  • Strikeout%
    • What matters? Velocity, only greater than 79 mph. The faster the better.
  • Ground ball%
    • What matters? Velocity, the faster the better until just over 80 mph. No clear advantage to throwing harder than 80.

Posted by Jeb

From the great "vacationland" state of Maine. Former D3 baseball player on an underachieving team. Prior: TrackMan with A's. Current: Check Down Sports. Soon: Video with Reds. All-time facial hair lover.