from Amazon
Monday, December 01, 2008
The 3-0 count
I love all articles on the 3-0 count because the number of choices to the pitcher and batter are very limited, while the payoffs are more apparent. Here’s John Walsh:
I re-did the above, this time only considering cases when first base was occupied. The thinking there is that with first base occupied, the pitcher is much less likely to pitch around a batter. The data show that there was no difference with a runner on first; the ball percentage was still 35 percent.
Later John says using PITCHf/x:
What I’ve done is take all 3-0 pitches, and thrown away anything that is not classified as a fastball. I figure if a pitcher is throwing something other than the fastball, he’s probably afraid the batter might swing on 3-0 and he’s unlikely to be aiming down the middle.
Perhaps John can merge the two, and look at situations where 1B is not open AND the pitch is a fastball.
As usual from John, the article is very well written with a nice visual chart.
Nice comments so far on ballhype as well. I agree with MGL and other commenters there that John’s assumption that pitchers are aiming for the middle of the strike zone is not strongly supported by the supporting evidence which he gives - that the average location is roughly down the middle. [Without that assumption, his quantitative conclusion about accuracy does not follow.] I’ll just add that since batters actually do swing 5% of the time on 3-0, they are probably prepared to swing if they get a pitch they like somewhere around 10% of the time. That should be enough to force pitchers away from routinely throwing their best controlled pitch right in the middle of the strike zone. Also, taken together, John’s Pitchf/x charts and text imply that around 10% of pitches thrown are not fastballs anyway - more evidence that even the 3-0 count does not permit an automatic strike. On the other hand, if pitchers are aiming for the corners and the actual location of pitches just averages out to down the middle, I would expect to see some trace of secondary clustering of pitches in the corners of the strike zone in his graphic, and I can’t make out any such clustering.
Finally (using retrosheet), I failed to exactly reproduce John’s tables of the pitchers who threw the fewest and most balls on 3-0 count for 2006-07. For instance he has Brandon Webb as the 7th most accurate pitcher with 67 pitches on 3-0 and .254 (17) being balls. I have Webb as the most accurate according to retrosheet: he threw 78 pitches on 3-0 with 10 being intentional balls. Of the other 68, 10 were unintentional balls (53 called strikes, 4 put in play and 1 foul ball). As a curiosity, Webb was perfect throwing strikes to right-handers (18 called strikes and a foul).
The most inaccurate pitcher according to John’s cutoff of 50 pitches should be Carlos Marmol at .481 (25 unintentional balls and 27 called strikes, as well as 5 intentional balls), but Marmol does not appear in his bottom 10. Jamie Wright, whom John lists as the most inaccurate at .508 (30 balls in 59 pitches), I have at .467 (28 balls, 30 called strikes, a foul and a ball in play, as well as 6 intentional balls.)