Here's a Thought: Examining Fastball Effectiveness in Major League Baseball

Nathaniel StoltzSenior Analyst IJuly 12, 2009

ATLANTA - JUNE 27:  Starting pitcher Tim Wakefield #49 of the Boston Red Sox pitches against the Atlanta Braves at Turner Field on June 27, 2009 in Atlanta, Georgia.  (Photo by Kevin C. Cox/Getty Images)

A few days ago, I wrote this article, which looks at the fastball effectiveness of the Oakland Athletics' pitchers.

It was surprising to me that lefty Craig Breslow's 90-mph heater was the most effective, and that the hardest thrower, Santiago Casilla, had the third-least effective fastball of the group.

Intrigued by the results, I decided to take a look at fastballs in general, and see if velocity had anything to do with effectiveness.

I set up a spreadsheet (yep, here we go with the stats) of every pitcher who's thrown at least 20 innings in the majors this season. The spreadsheet contained fastball velocity, fastball usage (how often they throw it), and fastball effectiveness (as measured by Pitch Type Linear Weights).

With those three variables in place, I decided to see if any of them correlated to each other:

  • Do harder throwers throw more fastballs?
  • Do harder throwers throw more effective fastballs?
  • Do pitchers who throw more fastballs get more effectiveness from them?

Obviously I can't paste Excel graphs into this article, so I'll just summarize my results. If you'd like to see the graphs yourself, just let me know and I can email them to you.

The first thing I looked at was velocity and usage. I excluded pitchers that throw cutters from the data, because if a pitcher throws a cutter, they throw their regular fastball less, so that skews the data.

I did get a slight trend on that relationship. For you math nerds, the equation of the regression line is y=.011x-.416 and the R-squared value is .091.

For those of you who don't know or care about math (and trust me, outside of sports, I don't know or care either), that means that fastball velocity influences fastball usage by 9.1%, and we can expect usage to increase a bit with velocity.

For example, we can expect, with this formula, that a pitcher with an 85-mph fastball will throw it 51.9% of the time. A pitcher who throws 100 mph can be expected to use the fastball 68% of the time.

Of course, that's far from a strict rule: for example, the slowest thrower on the list, Brad Ziegler, uses his fastball 78.3% of the time, and he only throws 84.3 mph.

So to sum up that relationship, yes, harder throwers throw more fastballs, but not to the degree one would expect.

The next thing I investigated was velocity and effectiveness.

If the previous graph's lack of correlation wasn't surprising, this one's sure was.

The regression line I got was y=.037x-3.53. That seemed fair enough.

The line also has an x-intercept of about 94 mph. That means that the velocity needed to expect average effectiveness is about 94 mph.

That's all well and good, but the R-squared value of the graph is .007. That means .7% of fastball effectiveness can be explained by velocity.

.7 percent R-squared basically screams "These variables have no correlation."


Apparently, fastball velocity has ABSOLUTELY NOTHING TO DO with fastball effectiveness.


So, with that long-standing belief seemingly shattered, I decided to see if fastball usage had anything to do with fastball effectiveness.

Again, for this relationship, I took the cutter pitchers out of the sample.

And once again, the graph is just a big blob of haphazard dots.

The equation for the regression line is y=1.678x-1.152 and the R-squared value is a measly .018. Fastball usage is responsible for 1.8 percent of fastball effectiveness. Again, that's basically random.

So what have we learned?

Well, the next time you hear your favorite team has a pitching prospect who throws 96 mph, don't get too excited, because the velocity on the pitch has nothing to do with how well it works. Joel Zumaya's fastball is nearly 2 mph faster than that of any other pitcher. It rates 221st out of 376 pitchers in effectiveness.

As for what actually makes a fastball good, there's a bunch of other factors, some more quantifiable than others.

Deception, command, and movement may exert more influence over fastball effectiveness than velocity or usage, but as I don't have that data, I can't examine the relationship right now.

Further proof that velocity isn't everything is the man pictured at the beginning of the article.

Oh, you thought picking Tim Wakefield was just ironic for an article about fastballs?


He's got the second-most effective fastball in the majors this year, behind Arizona's Dan Haren.

So, in conclusion, don't get too worked up over a pitcher's fastball velocity, whether it's 80 or 100. It has nothing to do with how effective the pitch is.

I'm interested to see if velocity and usage impact other pitches, so I'll tackle that in another article.