Wakefield’s Fastball Redux

A couple weeks ago Other Dave noticed that Tim Wakefield has one of the best fastballs so far this year. He suggested that Wakefield’s fastball is so successful, despite working in the low-70s with average movement, because it is a good 7 or 8 mph faster than his knuckleball and keeps hitters off balance. I really liked this idea and wanted to see if Dave was correct.

So I went through and looked for at-bats in which Wakefield threw a fastball after throwing at least one knuckleball in that at-bat, and found the difference in speed between that fastball and the knuckleball that immediately preceded it. First let’s look at the run value of a fastball based on its speed, the black line is the average and the gray standard errors. The run value is the change in run expectancy after the pitch, so a negative number is good for Wakefield.


To begin with notice that his fastball is quite good, -0.02 runs per pitch is -2 runs over 100 pitches, which is great. Interestingly after Wakefield’s fastball gets up around 72 mph there is no increase in effectiveness with an increase in speed. This is pretty surprising, generally the faster a fastball the better the outcome. Now let’s look at the run value of a fastball based on how much faster it was than the preceding knuckleball.


Here you see a clear consistent, if noisy, trend. As the fastball gets faster compared to the previous knuckleball its success increases. These two graphs together tell us that it is not the absolute speed of Wakefield’s fastball that determines its success, but its speed relative to the previous knuckleball.

Just as Dave suggested the success of Wakefield’s fastball is indeed tied to how much faster it is than his knuckleball, and since his knucleball is so slow he can be effective with his low-70s fastball.

Print This Post

Dave Allen's other baseball work can be found at Baseball Analysts.

Sort by:   newest | oldest | most voted
Big Oil
Big Oil

Very cool analysis; I don’t know how possible it would be, but a similar breakdown, refined along the lines of his knuckleball being “on” or “off”, would shed an interesting light on the fastball success vis a vis movement on the slower knuckler. I suppose you’d have to have some sort of movement criteria, perhaps derived from pitch f/x, to do so.