How Important is it for Pitchers to Add and Subtract on their Fastballs?

It is often said, or at least i hear it especially on television broadcasts, that a pitcher "adds and subtracts" from his fastball. That is, the pitcher will throw his fastball intentionally at different speeds. This is said to be a good thing, which makes some sense. One could think of situations where a hitter knows that a pitcher's fastball is usually 92 MPH, and instead of going to a different pitch, the pitcher throws an 89 MPH fastball. Or, Justin Verlander is often credited for being able to throw 100 MPH, but sitting at a lower velocity and maxing out when he needs to (at the end of his outing, or with runners on base, etc.). I have really had no idea how to treat this aspect of pitching other than just using averages, so I wanted to see if I could test whether or not "adding and subtracting" to the fastball was in fact a good thing. Perhaps, one could imagine, consistency is good, as it shows that a pitcher can maintain his velocity.

Since pitchers' average fastball velocities can fluctuate from year to year (from start to start as well, but not as much) it probably isn't wise to look at multiple seasons, so I will look at just 2013 so far for this article. This gives us small sample size problems for the season when trying to grade performance, so we will look at the swinging strike percentage of the fastball, along with the FIP - of the season.

In order to measure how much a pitcher "adds and subtracts" from their fastball, I'll use the standard deviation of all 4-seam fastballs (FF designation), which should tell us the difference on average for each fastball, which I think should give us an idea of who "adds and subtracts" the most. Of course, we are relying on MLBAM tags, so there may be some problems, as it seems that slower than usual fastballs by pitchers are usually called changeups. I don't know of a way to sidestep this problem. I looked at pitchers with a minimum of 1000 pitches thrown (regardless of classification) and at least 30% of pitches being the 4-seamer to weed out relievers and give us a look at starters with a decent sample size who use the 4-seam fastball with some regularity.

According to MLBAM, Bartolo Colon wins the award of being able to add and subtract to his fastball. Jason Hammel adds and subtracts from his fastball the least. The average standard deviation is 1.35 for the selected pitchers.

The pitchers with an average ST or higher had a whiff percentage of 6.04, the ones with a below average ST had a 5.92 whiff percentage, not a real difference. The below average ST pitchers had a 99.36 FIP - on average, with the average to above average having a FIP - of 96.75, again slightly better, but not a big difference. The top 10 do a little better when it comes to FIP -, but do worse when it comes to swinging strikes. The ten worst ST pitchers have a pretty significantly worse FIP -, with a below average whiff percentage, but one that is still better than the top 10 ST pitchers.

I think what helps show that adding and subtracting doesn't seem to be a determiner of success either way is that there are three pitchers with whiff percentages on their fastball over 10% (Shelby Miller, Matt Harvey, and Julio Teheran) and one of them have an above average ST, one below, and one about average. The thing they do have in common though is velocity, being in the elite part of starters.

Just plain velocity is a much better indicator of fastball success (though the high whiff percentages of Hisashi Iwakuma and Jose Quintana are certainly interesting) than adding and subtracting velocity, which doesn't seem to help much, but doesn't really hurt either.