THE BOOK--Playing The Percentages In Baseball

Of course one of the things that stand out in these graphics is the great command he has.  However, we really need to see other pitchers, or all pitchers, in order to put that into context. I mean if we look at the iron shots of the worst PGA golfer on the tour, it might look like he has amazing accuracy, but if we compare him to other PGA players…

I’m not saying that Mo does not have great command even as compared to other MLB pitchers - I would just like to see how much better rather than marvel at how good his command is compared to...I don’t know.

Also, any RH pitcher who throws a cutter is going to throw primarily outside to a RHB and inside to a LHB so as not to accidentally leave that cutter in the middle of the plate.  Plus you want them to chase cutters that are not in the strike zone. If you throw inside cutters to RHB, they won’t swing at pitches not in the zone.  Same for outside cutters to LHB (again, if you are a RHP).

Also, without knowing the game situations (or even the type of batter), looking at location by count is not very instructive, since the graphic is a composite of all situations.  For example, when the batter is the tying or winning run (and he has any power at all), most pitchers like to pitch primarily outside (and low), regardless of the count, so they don’t give up the HR even on an bad pitch.  But if the batter is NOT the tying or winning run, they will use all parts of the plate, regardless of the count.  Now, Mo is not like that because he has such good command (when he pitches inside, he does not miss “middle in” too often, so he does not have to fear the HR as much as most pitchers do), however, he still probably pitches differently depending on the batter and the game situation.

Wouldn’t you run into some pretty big sample issues if you break it down to that level, MGL?

Maybe and maybe not.  For a pitcher like Mo, who is very accurate with his pitches, you eliminate a lot of the noise associated with missing the target.  Then you merely have the issue of his pitch selection/location being random.  But since he throws mostly one pitch, you pretty much eliminate the pitch selection issue.  As far as location, it depends on how much he varies it.  For example, if in certain situations he throws inside cutters to LHB 90% of the time, sample size is not going to be much of an issue.  I would start with is the data and go from there.

The point I was trying to make is that sometimes (not always) looking at average results obscures the chance to garner any meaningful information.  That is especially true with pitch selection and location, although certainly looking at a pitcher’s pitch selection/location distribution across several different situations is a good starting point.  What makes it really necessary to look at pitching in the context of game situations (mostly outs, score and runners) is that it is so critical to pitch differently in different situations in order to be effective. For example, pitchers pitch completely differently (or at least they should) in close games with 2 outs and runners in scoring position, especially with base(s) open, then they do in the same close games with no outs and no one on.  And of course pitchers pitch completely differently to different batters.