Debunking!

By Jason Wojciechowski on August 2, 2003 at 7:49 PM

In this Oakland Tribune article, Carl Steward says

Even players who do not have a lot of walks like [Derek] Jeter and [Bernie] Williams still work the count, and the pitches can add up in a hurry.

Let's see what we can do with that statement.

First, Bernie Williams certainly does have a lot of walks. His 39 may not seem like a great number, but that's come in less than 300 plate appearances. He draws a walk every 7.7 plate appearances. If he had enough PA's to qualify, that would slot him 31st in the majors.

Jeter, on the other hand, he's right about. He walks once every 11 PA's or so, though that's not really all that bad. That's actually a little worse than his career rate, which is much closer to 1-in-10 than I ever would have guessed it was. 1999 was his big year, as he walked 91 times in 739 PA's en route to an MVP-type .990 OPS. Of course, that season preceded three seasons of decline, bottoming out last year with a .794 OPS, fueled largely by his worst batting average and slugging percentages since 1997. He's back to his career norm this year, though. But I guess the point is that, no, Jeter doesn't draw a ton of walks, so Steward got that right.

What about those "work the count" claims? Jeter sees 3.79 pitches per plate appearance, which would rank him about 77th in baseball among qualified players (of which, according to ESPN, there are 169), so he's just above the mean, so he doesn't work the count significantly more than any other major league hitter does.

Williams we'd expect to see a lot of pitches, since Steward was wrong that he doesn't walk much. This turns out to be surprisingly untrue, since Williams' 3.80 P/PA mark is only slightly above Jeter's mark. That's an interesting finding. I wonder what the P/PA looks like for most people with walk totals like Bernie's?

The ten players immediately surrounding Williams in the BB/PA chart who have enough plate appearances to qualify are

Name BB/PA
Jim Edmonds .133
Ryan Klesko .133
Trot Nixon .132
Carlos Beltran .131
John Olerud .131
Bernie Williams .130
Troy Glaus .125
Corey Koskie .125
Sammy Sosa .125
Luis Gonzalez .123
Austin Kearns .121

Now let's add a third column, that shows pitches per plate appearance, and see what happens to Bernie.

Name BB/PA P/PA
Jim Edmonds .133 3.95
Ryan Klesko .133 3.74
Trot Nixon .132 3.77
Carlos Beltran .131 3.81
John Olerud .131 3.88
Bernie Williams .130 3.80
Troy Glaus .125 4.03
Corey Koskie .125 4.00
Sammy Sosa .125 4.09
Luis Gonzalez .123 3.84
Austin Kearns .121 3.97

So we've got Klesko and Nixon below Bernie despite leading him in walk rate, Beltran about the same, with essentially the same rate, and everyone else ahead, with most of them being far ahead except for Gonzalez, who has only a slight edge. So at a quick glance, it looks like it's slightly unusual for Williams to see only that many pitches and get a lot of walks anyway. I wonder if this might indicate a possible future trend of less walks.

Evidence of past years says this might be the case, since his pitches per plate appearance is about where it always is, but his walk rate is the highest it's been since his exceptional 1999. As a matter of fact, that 1999 season is the only one in which he posted a better walk rate than he's working with this year.
So maybe Bernie's been a teeny bit fluky this year. That said, regressing to his mean is going to mean losing no more than a couple walks for the rest of the year, probably like two or three at the most, maybe less, so it's not going to affect his value in any significant way.

If it weren't so late, I'd do the reverse: pull up his comparables in terms of pitches per plate appearance and compare their walk rates. Better would be a straight correlation study between pitches per plate appearance and walk rate.
Obviously, they should be highly correlated, but I wonder how highly.