THE BOOK--Playing The Percentages In Baseball

Friday, October 02, 2009

Pitch sequencing with Mariano Rivera

Great stuff from Dave.  The reason that Mo provides such a great study is that he only has one pitch type, so the only thing we have to worry about is location.  So, Dave gives us the payoff.  First, he tells us how frequently he pitches inside, given that he’s following a previous inside pitch or a previous outside pitch:

vLHB vRHB
all 0.58 0.45
following inside pitch 0.63 0.55
following outside pitch 0.40 0.37

Now, those numbers, by themselves, tell us nothing.  We don’t know how often he should pitch inside, if the previous pitch was an inside pitch.  Is 63% good?  Should it be 50%?  75%?  Exactly what is the batter thinking it should be?

Well, Dave does the right thing by telling us the run value of that next pitch.  Here it is against LHH, as run values per 100 pitches (roughly a 6 inning game):

rv100 vLHBs
inside outside
following inside pitch -3.4 -3.9
following outside pitch -2.3 -2.8

We see that his outside pitches (the more negative the better) have more value than his inside pitches.  Not much more (and Dave said it’s not stastically significant).  To the extent that the differences are real, then Mo should throw a bit more to the outside than he does (whether the previous pitch was inside or outside) against LHH.  Since he throws an inside pitch 63% of the time when the previous pitch was inside, maybe he should cut that down to 60% (just guessing).  And if he throws inside 40% of the time following an outside pitch, maybe he should bring that down to 37% or something.

I should note that it is CRITICAL to figure out the talent level of the batters at this point.  It’s all well-and-fine to show the values Dave shows, but we need to know what kind of hitters he’s throwing at.  If for example he throws disproportionately inside to good hitters, then that’s why the run values on inside pitches don’t look as good as outside pitches. 

But, this is blow-my-mind numbers:

rv100 vRHBs
inside outside
following inside pitch -1.7 -0.4
following outside pitch 1.4 -2.7

That is a huge difference, and likely impossible to link to the pool of batters.  Focus on the last line.  When he throws outside, when he goes outside-outside, he’s a monster.  When he throws outside-inside, he’s terrible.  He needs to cut down his outside-inside sequencing.  In his first table, Dave shows that outside-inside against RHH happens 37% of the time (that is, given that the previous pitch was outside, he will throw inside 37% of the time).  That number is way too high.

HOWEVER, say that he drops that down to 20% or 15%.  Guess what happens?  His run value will improve from +1.4 runs to, say -0.2 runs.  But, his -2.7 will worsen to say -1.7.  Or something.  Whatever.  Therefore, it’s very easy for me to say that he needs to cut that his outside-inside sequencing to RHH.  But, we also need to know what happens to the effectiveness of his outside-outside sequencing.  His tremendous results on the outside-outside may be directly caused by his horrible outside-inside, as batters look for outside-inside a bit more because they know Mo is vulnerable there, and therefore, leave outside-outside for Mo to exploit.

This is the most beautiful part of baseball: game theory between batter and pitcher.