THE BOOK--Playing The Percentages In Baseball

----"However, more important than the “at” pitch count is the “through” pitch count”

I’ve seen this same claim from others, such as Tippett and Palmer, I think, and so I always assumed it was true.

But let’s say a batter has a 3-2 count. Here are two pathways he could have used to arrive there: A) 0-0, 1-0, 2-0, 3-0, 3-1, 3-2, B) 0-0, 0-1, 0-2, 1-2, 2-2, 3-2.

Using the “through” approach, batter A will get lots of positive credit, and B will get negative credit. But both players have taken the same number of balls and strikes, and both are in the same position with regard to concluding the PA.

Looking at Tango’s example of the 3-0 count--just because it’s not surprising that almost all PA that end at 3-0 are BB, doesn’t mean that it is not important.  And if it instead goes to 3-1 and ends there, the batter will get credit for ending a PA at 3-1 (likely to walk or get a good pitch to hit).

Using the “through” approach, batter A will get lots of positive credit, and B will get negative credit.

Dave - I don’t understand what this sentence could possibly mean.  I’ve not seen anyone propose a batting metric that gives “credit” for passing through a particular count.  Could you give examples?

For me, count information is useful in several ways.  BABIP at different ending counts is useful in demonstrating how much control pitchers have on outcomes that were previously thought to be mostly in the control of the batter and fielders.  Through counts are useful for studying different approaches to batting and pitching and how the count can influence certain strategic decisions.

Peter, such a metric could easily be constructed as a way to evaluate a batter’s approach. You could take the final count of each of a batter’s PAs and apply an average result for that count, instead of the batter’s actual result. Sort of like the stat called PrOPS, where they do the same thing to batted ball types, IIRC.

I do agree with your second paragraph.

My point was simply that the “at” counts are also valuable info, depending on what you are trying to do.

David, I reject the conclusion:

“Using the “through” approach, batter A will get lots of positive credit, and B will get negative credit.”

They will get identical credit.  Or should anyway.  If the linear weights by count that people have published are showing otherwise, then let’s have that discussion.

How will they get identical credit? Player A has uniquely gone thru positive counts of 1-0, 2-0, 3-0, and 3-1. Player B has gone thru counts of 0-1, 0-2, 2-2, 1-2, and 2-2.

Why would they get identical credit? What am I missing?

The negative counts in #5 were messed up. I meant 0-1, 0-2, 1-2, 2-2.

Going from 3-0 to 3-1 is negative, as is going 3-1 to 3-2.

But I’m not talking about the transition, just about whether the batter passed thru a certain count. 3-0 and 3-1 are both positive.

David, you said:

“batter A will get lots of positive credit, and B will get negative credit”

The “credit” is based on the transition.  And it is irrelevant when you go from 0-0 to 3-2 how you get to 3-2. He’ll get the same overall credit.

By “credit” I didn’t mean the transition--I meant positive credit each time a batter is in a hitter’s count. You can get this from a source like B-Ref in their splits. You need lots more data to get each pitch to pitch transition