A Surprising Benefit of Throwing a Good Sinker

*Note: all stats are as of August 1, 2017

I originally intended to write a post about the aspects of a four-seam fastball that are most important in generating whiffs. The correlation between fastball velocity and whiff rate on the fastball is only about 25%, so I was interested to find out whether other factors, such as vertical movement, location, or pitch usage, are better indicators of a fastball’s swing-and-miss tendencies. While some of the names at the top of the fastball whiff list were not surprising at all (Chris Sale, Jacob deGrom, James Paxton), there were several others who I was surprised to see, including Brandon McCarthy, Rick Porcello, J.A. Happ, and Clayton Richard. There was one glaring similarity between these seemingly overachieving pitchers: they throw a high percentage of sinkers.

So I looked at the correlation between whiff rate on the four-seam fastball and sinker usage, only to find that it was not only small, but also negative. However, looking at the correlation between these two variables is somewhat like a chicken-and-egg problem: does sinker usage affect a pitcher’s four-seam fastball whiff rate, or does his four-seam fastball whiff rate affect his sinker usage? The latter option certainly seems reasonable: a pitcher who is ineffective with his four-seamer is more likely to develop a sinker than a pitcher with a dominant four-seamer. For this reason, we have to dig deeper to determine if sinker usage has any effect on four-seam whiff rate.

I looked instead at only the 48 qualifying pitchers who throw a sinker at least 10% of the time (and a four-seam fastball at least 5% of the time). I found the correlation between several variables — some relating to the sinker and some unrelated to it — and four-seam whiff rate. If the variables related to the sinker have a significant correlation with four-seam whiff rate, then that implies that a pitcher’s sinker can have an effect on his four-seam fastball. The variables I looked at were the four-seam fastball’s velocity and vertical movement, the sinker’s velocity and vertical movement, and the difference between a pitcher’s four-seam fastball and sinker in both velocity and vertical movement. Here are their correlations with four-seam fastball whiff rate:

	4-Seam Fastball	Sinker	Difference
Velocity	0.3022	0.2249	0.3011
V-Movement	-0.0544	-0.2875	0.4348

 

There are a few interesting things to note here. First, the four-seam fastball’s velocity seems to be just as important as the difference in velocity between the four-seamer and the sinker. While velocity is often the first thing most people look for to determine if a pitcher has a swing-and-miss fastball, relative velocity is equally as important as absolute velocity, at least when it comes to pitchers who also throw a sinker. This confirms the notion that changing speeds can upset the hitter’s timing and make a fastball seem faster than it is.

Relativity is even more important when it comes to vertical movement. While there is no correlation between four-seam whiff rate and four-seam vertical movement, there is a significant correlation between four-seam whiff rate and the difference in vertical movement between the four-seamer and the sinker (I’ll call this “v-movement difference”). This seems to show that the downward movement of the sinker makes hitters more likely to swing under the four-seam fastball; they keep the sinker in mind, so the four-seamer appears to have more vertical movement than it actually does. If this is true, then we should expect v-movement difference to have a greater effect on pitchers who throw a higher percentage of sinkers. To test whether this is true, I increased the requirement of minimum percentage of sinkers thrown in intervals of 5%, from 10% to 35%. I then found the correlation between four-seam whiff rate and v-movement difference at these different thresholds. Here are the results:

threshold	correlation
10 %	0.4348
15 %	0.4125
20 %	0.4752
25 %	0.4752
30 %	0.5121
35 %	0.5025

 

Just as we expected, the correlation between v-movement difference and four-seam whiff rate is higher for pitchers who throw more sinkers. If the relatively high correlation we observed at the 10% threshold were pure luck, then the correlations at higher thresholds would be scattered randomly. The fact that there is a clear upward trend in correlations as the threshold increases proves that v-movement difference does, in fact, have an effect on four-seam whiff rate. While this does not necessarily mean that adding a sinker will help a pitcher get more whiffs on his fastball, it does prove that the quality of a pitcher’s sinker can affect the effectiveness of his fastball. More specifically, we also learn that a good sinker, in terms of generating whiffs on the four-seamer, is one that has little vertical movement (or a lot of sink) in relation to the four-seamer.