Outside Effects on the Knuckleball
The knuckleball has always been an interesting pitch to watch and study. It has the most interesting physics of any pitch in baseball, with the possible exception of the mysterious gyroball.
With the inception of Pitch F/x data, we have been able to study the knuckleball in ways never even imagined. It allows us to answer many questions about the effects of outside elements, such as the weather, on the knuckleball.
Before I answer any questions, I feel I should quickly explain my methods. I obtained all of Tim Wakefield's 2008 data, and broke down all of his knuckleballs. Because it would be difficult to measure horizontal and vertical movement together with all of the different factors, I used the Pythagorean formula (a^2 + b^2 = c^2) to compute how much the pitch moved from the zero mark. I called this "movement", and it appears on all graphs but the first one.
First, let's look at Wakefield's knuckleball movement throughout the year.
The red circle on the graph represents the average movement of every knuckleball.
Next, let's look at how temperature and wind effect the movement of the knuckleball.
As it says on the graphs, there isn't much of a correlation between movement and the temperature and pitch movement. A correlation of 0.21 is small enough to conclude that the temperature doesn't have much, if anything at all, to do with knuckleball movement.
However, wind had more to do with the movement. A correlation of -0.375 tells us that with less wind, there is more ball movement. However, the correlation isn't big enough to conclude that wind and movement are inversely correlated or not. It seems to be contrary to popular belief though, because you would think that faster wind would move the ball more.
Wakefield played games in domed ballparks last season, giving us a look into how his knuckleball moved in a more neutral environment. Here is how his indoor and outdoor games compared last season.
Strange results. Wakefield's outdoor results are more clustered, while his indoor games are spread out. Once again, this goes against what one would expect. Given zero wind and a constant temperature, I would expect to see the indoor games in a tight group, and the outdoor games in a spread. Granted, the indoor games had a smaller sample size, but it still jumps out to me.
Lastly, let's look at the effect of the speed of the pitch and the movement.
Once again, a correlation of -0.15 tells us that speed and movement really have nothing to do with each other. In fact, the faster the ball moves, the less it will break.
The knuckleball still remains a mystery due to the lack of consistency of Pitch F/x to correctly measure a knuckleball, and the weirdness of the pitch itself. All the above data proves is that the knuckleball will do what it will do, and there is not much anyone can do about it.
Thanks to Brooks Baseball for making the .xml data easy to find and download, and thanks to Sky for the idea for this piece.
UPDATE: Thanks to an astute observation, I took out all pitches that appeared to be fastballs. As predicted, the data didn't change a whole lot, but it didn't make a difference.