Bill James Guide to Baseball Managers, The (40 page)

The theory of this book, as you may remember from the introduction, was to ask very specific questions about baseball managers and generate specific, objective answers to those questions.

Some years ago, in writing
a Baseball Abstract,
I was struck by the fact that the discussion of baseball managers proceeds almost in a vacuum of organized information. Take two baseball players—let’s say Jay Buhner and Steve Finley. If you ask a baseball fan a series of comparative questions about those two, like “Which one has more power?” “Which one runs faster?” and “Which one has a better throwing arm?” he will know the answers. Even if the difference between the two is very slight, like the difference between Cecil Fielder’s speed and Harold Baines’s, he will still know.

But if you ask parallel questions about two
managers,
he will
not
know the answers. Who is quicker to go to the bullpen: Don Baylor or Felipe Alou? Who platoons more: Bobby Cox or Tony LaRussa? Who uses the hit and run more: Bob Boone or Phil Garner? The average baseball fan, unless he has read this book, doesn’t have a clue what the answers to these questions would be.

Think about it this way: What do you put on the back of a manager’s baseball card? Take one and flip it over. We have hundreds of statistics about baseball hitters, but we know how to summarize a player’s abilities into a simple chart which tells us what the hitter does well, and what he doesn’t do so well. Same thing with a pitcher’s card—we condense his performance into ten or twelve categories which tell you how much he pitched, how well he pitched, and a little bit about what type of pitcher he is. You look at the back of a manager’s card, and it will tell you that he hit .238 at Pittsfield in 1971. It doesn’t have anything to do with how he manages.

What,
specifically,
does one manager do that another manager might not do? Does he pinch-hit a lot, and if so, under what conditions? Does he like to platoon? Is he aggressive about using his bullpen? Does he hit and run much? Is he prone to bring the infield in? Does he favor the intentional walk? Simple, objective questions which have simple, objective answers.

I began addressing these subjects about ten years ago. We haven’t made great progress in the last ten years, but the discussion has advanced a little bit. Old records have been compiled which will tell us, for example, how many relievers each manager used each season. The wizards of STATS Inc. have begun collecting specific information about many related issues, such as hit-and-run plays and attempted bunts. We have a good deal of information now that we did not have ten years ago.

But the discussion has not moved forward as much as perhaps it should have because that information has not yet reached the public. It hasn’t reached the public because it has not yet been put into
a form
that the average fan can use.

That is the purpose of this article:
to construct what might become, with luck, a “standard form” of a manager’s record.
How do we put this information into a form that the customer can use? In a record book we have a form for batters’ records, a form for pitchers’ records, a form for fielders’ records. What is the form for managers’ records?

Baseball cards can have anywhere from eight to eighteen categories in the back-of-the-card record. Media guides usually have fourteen to sixteen categories in the batting or pitching record. Anything above 15 is pushing our luck, plus, if we do this job right and people actually start to
use
the form that we establish, there will be pressure to add things. People will start to say, “Well, I think the record ought to have included this or that or the other.” When they say that often enough we’ll have to accommodate them.

As a starting point, we’ll need to have “games managed.” Enough managers are fired during the season that, if you don’t have the games to establish context, you’ll get confusing data, one guy with 128 games managed mixed in a chart with the other guys at 162 games. Also, off to the right, we’ll need the won-lost record and the winning percentage—just the basic information which is in the Managers Register of the
Baseball Encyclopedia.
Let’s do Tony LaRussa, since he’s a great manager and I haven’t said much about him in the book:

LaRussa’s career won-lost record is 1,408–1,257. I’m not going to run a totals line, because as this record develops we’ll have many blank spaces for missing data.

There are a number of other things that we
could
put in here. The Macmillan
Baseball Encyclopedia
has a column for “ties” and a column for “no decision,” which is obviously a waste of space, because you’ve got a whole lot of zeroes and ones there.

The
Baseball Encyclopedia
also has a column for “standing,” for which one can make a better argument. This column says “3” if the team finished third in their league or division. If the manager was fired with the team in fifth place and they improved to finish fourth, there would be two numbers in the column, a “5” in boldface, and a “4” which is not in boldface.

If I was designing this chart thirty years ago, I would probably have gone with that, but I’m not going to include it here. In modern baseball the number of possible “positions” is not consistent. When you see a “4,” you don’t know whether that’s fourth in a four-team division, or fourth in a five-team division. If it was a record from 1991, it could be fourth in a seven-team division. You wouldn’t know instinctively, when you saw the number, what it meant. In modern baseball, a second-place team can win the World Series, so if you put in this category, you’d have to have a code saying “WC,” for “wild card.”

To some degree you can replace the information by using boldface. If a team finishes 91–71, the baseball fan knows automatically that that’s a good record. If the team finishes first, you can put the “91” in boldface. That’s what the reader needs to know. If the team finishes 75–87, the reader doesn’t really care very much whether they finished third or fourth.

Let me suggest this. If the team finishes first in the division, we’ll put the wins in boldface. If they finish with the best record in the
league,
we’ll put both the wins and the winning percentage in boldface. Okay?

A couple of other things that we
could
do here. We could put in the performance of the team
against their expected wins.
Remember that formula—expected winning percentage equals:

Last year’s winning percentage times .5, plus
The previous year’s winning percentage times .125, plus
The year before that’s winning percentage times .125, plus
A .500 record times .25.

We could figure the expected wins for each team, and thus highlight the good years in the manager’s record by showing them as “+11” or whatever, while the weak seasons would be shown as minuses.

I’d like to do that, and I’d choose that over the “team standing” column, but I can’t quite recommend it. First, it is different from the information which is traditionally included in a player’s record. We don’t have a category next to batting average which tells you what the player’s expected batting average was. This
is
a different situation, so I could get beyond that argument, but there is also the matter of explaining it to the audience. A record is supposed to be automatically assimilated by the reader. Putting in “versus expected wins,” or “VEW” for short, would raise a question in the reader’s mind. Expected versus what? Who says St. Louis was expected to win 79 games in 1996?

Or, on the other hand, you could put “+12” there when the team
improved
by 12 games, and “-7” when the team declined by 7 games. This presents its own problems, however. Suppose that the team went 83–79 last year, started out 41–56 this, and that manager was fired. The team then went 30–35 under the new manager. What do you put there? They were better than they were under the previous manager, but still not as good as they were last year. Do they get a “+” or a “-”?

You could use “games behind”; that’s kind of standard, too. But again, who cares? If the team finished 78–84, do you
care
whether they finished 12 games behind or 14 games behind? Perhaps, but we’ve got limited space.

So let’s not do any of that. We’ve got wins, losses, and winning percentage; that’s all we have room for. From now on, let’s concentrate on describing the performance of the manager.

Well, what does the manager do?

The manager writes a lineup. There is a statistic which has become common in recent years, which is
the number of distinct lineups used by the manager during the season.
A few local newspapers are using it, and it does tell you something significant about the manager, which is whether he likes a settled lineup, or prefers to shuffle. Let’s do that for Tony LaRussa. I won’t do the back years, because I don’t want to make up data, but we have the information (from STATS Inc.) for the years 1993–1996:

The “149” is boldfaced for 1993 because that led the league.

Next to the number of lineups used in the STATS
Major League Handbook
are two categories entitled “% LHB vs. RHSP” and “% RHB vs. LHSP.” For 1996, the start of this chart looks like this:

This, to me, is a classic case where less information would be more. The problem with this chart, for me, is that it pushes my mind through a series of mental flip-flops before I can process it. Let’s see now; would a “platoon” manager be a manager who has a large
total
in these two columns, or a manager who has a large
differential
between the two?

What we want to know, I think, is
the percentage of starting players who had the platoon advantage.
If you give us that, we’ve got information that can be readily understood. Bob Boone has a “Pl%” (“platoon percentage”) of .741, meaning that
74.1% of the hitters in Bob Boone’s starting lineup have the platoon advantage.

If we give the reader just that, then standards and norms will develop rapidly. See, we are trying to develop information which everybody will understand, and everybody can use. If we put out the information in this form, then a TV announcer can say, in the first year, that “Phil Garner had the platoon advantage with 58.6% of the players in his lineup last year. That’s a little below the league average, which is 62%, so he doesn’t platoon a whole lot.” Then in the second year, the ESPN guy can say, “Phil Garner’s platoon percentage last was .586, while the league average was .621, so he’s not really a platoon manager.”

And then in the third year, the announcer can say, “Phil Garner’s platoon percentage last year was only .586,” and everybody will know what he means, because by that time the standard will be established in the viewer’s mind. He can say, with emphasis, “Bob Boone had a platoon percentage of
.721
last year,” and everybody will know immediately what that means. This is the data for Tony LaRussa, as best I know it: