THE BOOK--Playing The Percentages In Baseball

Great, great stuff.  I kinda wish he had included wOBA or at least OPS in those graphs, as we want to know about overall production.

There could be other reasons besides technology driving the changes over time.  For one thing, theories and approaches by batters and pitchers change over time for various reasons or for no reason at all.  It could be that the prevailing strategy in the old days was to be cautious and conservative the first time up, and now it is not.

Also, one thing that we didn’t address in The Book, and no one has mentioned that I know of is this:  The visiting team has a much larger gap between first and subsequent PA.  Much larger.  In fact, a good chunk of the overall HFA actually occurs in the first inning.  There are two possible reasons for this:  One, visiting batters have not yet played the field, so they are “cold” like a pinch hitter or DH (and thus have a similar “penalty").  Two, they are not yet familiar with the mound and lighting of the stadium.

It would be interesting to break down Matt’s numbers by home and road to see if perhaps it is the road team that is mostly shrinking the difference between some of the stats over time.  Maybe stadiums are more uniform now with respect to the lighting and pitcher’s mounds, or maybe players are more familiar now with opponent stadiums.

Anyway, great stuff. 

Also, It is probably not that necessary since we are only looking at the first two times at bat, but Matt, did you control for the pool of pitchers in your numbers?  The second time up is going to have SOME relief pitchers in it. Maybe there were more or fewer relief pitchers the second time at bat in the old days than now.  And if those (long) relief pitchers are worse overall than the starters, then you would see an artificial increase in batting stats from the first PA to the second, if you were not adjusting for the pool of pitchers in the numbers.  And maybe the difference in quality between starters and long relievers is different now than in the old days.

So I decided to only look at the first two plate appearances that a hitter gets against a starting pitcher

First sentence of paragraph 4.

FWIW, his numbers look similar to what I have in Table 81.

Peter, right, that was my mistake, and actually I didn’t mean to ask about that.  I meant rather, if he controlled for the pool of pitchers in the first and second PA, or did he only use PA’s where batters faced the starting pitcher at least 2 times?

I also forgot to say that from the pitch f/x research on pitcher stuff throughout a game, the difference we see in “times through the order” does not appear to have much to do with pitcher fatigue.

MGL: good point about pitcher fatigue.

MGL: if you look at Table… I think 82?, 83?, you will see an “expected” column.  AS you can see from that table, the pool of batter v pitchers remains roughly the same, such that we can presume the pools are equivalent, even without controlling for anything.

MGL: the difference in times through order is, I think, more fastballs are possibly thrown early and more breaking pitches late.  If someone can show fastballs thrown to each lineup slot, each time through the order, that may help.

Definitely more fastballs in the first inning, if I recall the pitch f/x articles correctly.  After that, I am not sure if there is much difference.  Plus, I think the fastball/off-speed ratio might be more of a function of inning than times though the order.

Again, I found the difference between the home and road team’s first inning fascinating.