Tim Lincecum tries to convince catcher Steve Holm that one more slider low and away will end a Phillies threat. Giants at Phillies May 4, 2008 (Icon/SMI) |

Runners are on first and second with one out and the burly first baseman is up. The count is 2-1 and he is sitting dead red when unexpectedly the pitcher unloads a curveball in the dirt and the slugger misses it by a foot. Now the count is 2-2 and the pitcher has a decision to make.

He knows the slugger has trouble with curveballs and could eat his 88 mph fastball for lunch, but he just threw a curve in the dirt. Wouldn’t the slugger be now be looking for another one? Maybe the slugger is thinking there is no way he comes back with another curve so that is what I should throw. But maybe the slugger thinks that the pitcher thinks that the slugger thinks… well you get the idea.

These are the kind of battles going on between the batter and the pitcher/catcher all the time in baseball and it manifests itself in the pitch sequence the pitcher uses. In the next few weeks I’d like to explore some pitch sequencing starting with the concept of “doubling up.” Doubling up is when a pitcher throws the same type of pitch in the same location as the previous pitch (whether on purpose or not). The question is how often pitchers do this and how effective it is. With a little help from PITCHf/x we will answer those questions.

###### Method

For this study, we will break the strike zone into nine regions, similar to what John Walsh did yesterday when looking at change-up location. (If you haven’t read this article, go do so.) This makes each region about 6.5 inches wide and seven inches tall. I am also going to include one region wide around the strike zone to make a nice five by five grid. If a pitcher throws the same pitch type in the same region two pitches in a row, that will be a double up.

I am also going to use John’s Runs100 for this analysis but I wanted a zero Runs100 to represent league average here so I normalized the data. I will call that nRuns100 and, like Runs100, negative is good for the pitcher and positive is good for the batter.

Any pitch of the same type that lands in the same bin as the previous pitch is considered to have doubled up. The strike zone is represented by the solid line. |

###### Results

So the obvious first question is how often pitchers do this. The answer turns out to be about 7.5 percent. That is interesting, but is that more likely or less likely than if a pitcher was just throwing his pitch types and locations at random? I binned each pitcher by type and location, then calculated how likely a double-up would be if he sequenced pitches at random. I came up with 10.7 percent. This means that pitchers are actively trying **not** to double up with pitches, though they still are doubling up a significant percentage.

Okay, so how effective is doubling up? Now we can use nRuns100 to quickly compare a doubled-up pitch to league average. It turns out that doubled-up pitches grade out at .99 nRuns100—almost exactly one more run per 100 pitches than league average. It appears that pitchers are correct to try to limit throwing the same pitch in the same location.

How do the different pitch types fare, you ask? Here is a breakdown:

type nRuns100 Fastball 1.35 Sinker 1.56 Curve 0.44 Slider 0.58 Change 0.15 Splitter 0.48 Cutter 0.34

Not too surprising to me—fastballs and sinkers grade out the worst. This may be a bit biased because pitchers will often throw a fastball on 3-0 and then again on 3-1. If they end up in the same bin it counts as a double up and while hitters are unlikely to be swinging at 3-0, they are ready to jump all over the fastball on 3-1. Off-speed pitches tend to hover around half a run worse than league average, with change-ups grading out quite well by comparison. Maybe hitters are thinking that because the pitcher threw a change-up last pitch, it is unlikely to see one again this next pitch.

How location-dependent is this? Can doubling up be useful as long as the pitcher keeps the ball away from a batter’s happy zone? We can break down pitches bin by bin to find out. Here, I am mirroring all data to left-handed hitters, so the low and inside bin means low and inside no matter which box the batter is occupying. Also, because of lack of statistics, I am going to list only the nine bins in the strike zone and only show the fastball distribution.

Inside Middle Outside Up -0.7 0.8 -1.2 Middle -1.1 2.9 -3.2 Down -0.8 -1.5 -2.4 All values are nRuns100 for fastballs

You can get some feel for how these numbers are different from standard fastball numbers by comparing them to John Walsh’s plot in his article looking at fastball speed. None of the bins do as well as what John found in his study. Keeping the fastball away seems to produce good results whether or not the batter has seen a similar pitch the pitch before, but if you try to go inside twice, or down the middle twice, you are just asking for trouble.

Interestingly, down the middle was the most often doubled up bin, checking in at 629 times. None of the bins outside the strike zone had enough statistics to draw any real conclusions except for the bin just above the Up/Middle bin, which would be fastballs shoulder high. Despite being out of the strike zone, that pitch checked in at a nRuns100 of 0.1, so if you really need a swing and a miss going up the ladder twice in a row probably isn’t the worst idea. All the other bins outside the strike zone were quite positive, as you would expect.

###### Conclusions

Pitchers generally try to avoid doubling up with good reason. Pitchers lose about a one run per hundred pitches compared to league average when doubling up on with a pitch. Fast balls appear slightly worse than off-speed pitches, but if thrown in the right location, fastballs can be effective. All locations, though, do worse than than just a single fastall thrown in that location. If you are lucky enough to get away with grooving one fastball down the middle, try very hard not to do it again because chances are the hitter isn’t going to miss twice.