Bill James Guide to Baseball Managers, The (41 page)

LaRussa’s platoon percentages are low, and the percentages tend to be higher in the American League than they are in the National; see, already we are learning something from doing this.

Another advantage of choosing these two pieces of information is that
this information can be re-created from box scores.
You could go back and figure, if you wanted to, how many lineups Casey Stengel used in 1955, and how many Paul Richards used, and how many Al Lopez used. You could figure out how many of those hitters had the platoon advantage, what the league norm was in 1955, and what it is now. In this way, you could get a better understanding of how the game has changed since then, as well as a better understanding of those managers.

Okay, we’ve got six categories of information in there now, leaving us nine to work with. Let’s do something about the handling of the pitching staff.

An obvious variable among managers is their tendency to use the bullpen, as opposed to their preference for leaving the starting pitcher in the game. There are many ways to measure this. We could, for example, put “complete games” and “saves” in the manager’s record.

There are a couple of reasons not to do this. Both complete games and saves reflect an element of ability, as opposed to managerial style. That isn’t a big deal, because after all, what doesn’t, but another reason not to include this information is that it is already in the record, in other places.

There is third reason, however, which is that
complete games are headed toward zero.
We’re within a few years now of complete game norms, for a team/season, below ten. That’s not going to be very helpful to the reader. Saves track wins, particularly if complete games are near zero, and we already have wins.

Another thing we could include would be “quick hooks” and “slow hooks”, which are categories that I invented about fifteen years ago, and which STATS figures and prints in the annual
Handbook.

That would be okay, but there are problems. You have to explain the categories to the audience. It’s a simple concept, but we don’t want to explain any more than we have to. A bigger problem is that the categories have tended to slip over the years, and will continue to do so in the future. I intended for them to “balance,” as many quick hooks as slow hooks, and they did balance fifteen years ago. There are a lot more quick hooks than slow hooks now, and if we moved the standards, we’d probably be back where we are in another fifteen years. I don’t think that’s what we want to do.

We could do something
like
complete games or “slow hooks,” which would be “long starts,” meaning any start in which the pitcher threw more than 115 pitches. That would be all right, but the best thing we can do, I believe, is just “relief games.” The information already exists, although you rarely see it. I think this is the one category here which best summarizes the tendency of the manager. Don Baylor in 1995 used 456 relievers; Dallas Green used only 298. Don Baylor used 449 relievers in 1996, leading the majors again, while the Mets under Dallas Green (for most of the season) once again used the fewest relievers of any National League team (325). That’s clear, it’s short, and it traces directly back to the manager. I think that’s what we should use:

Tony LaRussa

LaRussa never led the league in relievers used until his team was in decline in 1991, but he has often been among the league leaders. Again, I think standards would develop in this area, so that a TV announcer within three or four years would be able to say “Bucky Showalter used only 302 relievers last year,” and the audience would get it immediately.

STATS also prints games with three pitchers and two runs or less. The San Diego Padres in 1996 (under Bruce Bochy) had 32 games in which they used at least three pitchers, but gave up no more than two runs. That’s kind of redundant, I think. STATS prints “first batter platoon percentage,” meaning “When a reliever comes in the game, how often does he have the platoon advantage on the first batter?” This is an interesting concept, but the data are not as interesting as the idea behind them, since the answers tend to cluster between 58% and 64%. They also figure what we could call “long saves” or “LS,” which is saves of more than one inning pitched. That’s interesting data. On that chart, Lou Piniella and Felipe Alou stick out like an ex-wife at a family re-union, while some guys have almost no long saves … let’s put that on the “B” list, and we’ll work it in if we have room.

I think we have to try to describe, in this chart,
how the manager uses his bench.
Art Howe, in 1996, used 74 pinch runners. Jim Fregosi used only 12. That’s a real difference. It says something about what the manager values. Tom Kelly used 207 pinch hitters that year; Davy Johnson used 85. Joe Torre used 55 defensive substitutes; Cito Gaston used only 11.

This chart would be more interesting if we showed different managers, rather than different years for one manager, so let’s try that. We’ll add three categories: “PH” (“pinch hitters used”), “PR” (“pinch runners used”), and “DS” (“defensive substitutes”):

1996 Managers
National League

1996 Managers
American League

Most of this information, if you’re wondering, comes from the STATS
Handbook.

There are some notes worth making about this chart. Bobby Cox, the hugely successful manager of the Atlanta Braves, uses a very settled lineup, and Mike Hargrove of Cleveland uses one of the most set lineups in the American League. There is a great danger here of confusing cause and effect, but let’s move on. Jim Fregosi, who used defensive substitutes with great effect in his 1993 championship season, had lost confidence in his bench by 1996, and had more or less stopped making any in-game use of his reserves. Phil Garner rarely pinch-hits; he uses his bench for pinch runners and defensive substitutes. I did not know these things.

The National League, of course, uses more pinch hitters, because they don’t have the DH rule. Platoon percentages are higher in the American League than they are in the National because STATS includes the starting pitcher in the counts of how many right-handed hitters are in the lineup against lefties and vice versa. This is an illogical decision, I know, but they didn’t ask me. Even adjusting for that, there is somewhat more platooning in the American League, because the DH rule in effect expands the roster, since teams don’t have to carry as many pinch hitters.

You might have guessed that the DH rule would mean
less
substitution on defense, because the rule allows you to play a bad fielder without putting him in the field. In fact, the opposite is true: The American League uses about 35%
more
defensive substitutes, every year.

We’re going to have to pass up about fifty categories of information that we
could
add to the chart here. We could, for example, include pinch-hitting batting average, pinch-hitting home runs, and some code about whom the manager tends to pinch-hit
for.
We could include information about which position defensive substitutes are most often used at, who is pinch-run for, how many bases the pinch runners steal, how many runs the pinch runners score, how often a manager pinch-hits when he is ahead, and how often he pinch-hits when he is behind … we could make a monster chart out of how the manager uses his bench. The goal here is to focus on the most valuable information and reduce it to a cohesive table.

I want to develop a record here such that a general manager, reviewing the records of a half-dozen candidates to be his new field manager, could associate the tendencies of the manager with the attributes of his ball club. “Oh, I see,” he might say, reviewing the record of Phil Garner. “His bench is speed and defense. Well, I don’t know about that for our club, but he does vary the lineup quite a bit. That would be good for us, get some of the guys on the bench in the lineup a little more.” How would Terry Collins fit our team? Is Sparky Anderson worth another look? These questions need information. The form would summarize the information.

We’ve used ten columns so far; we have five left. I’d like to do something which describes the offensive/defensive orientation of the manager, but I don’t see how we can do it. Dallas Adams developed something called the “offensive fraction.” If a team scores a lot of runs and allows a lot of runs, or has a tendency in that direction, their offensive fraction will be high, over .500. If they tend to score not too many runs and allow fewer, attempting to win games 2–1, that creates a low offensive fraction, under .500.

That would be a good thing to include, but I don’t think I will. The offensive fraction is heavily indebted to the characteristics of the park. If a team plays in a hitter’s park, like Fenway or Coors Field, they’re going to have a high offensive fraction 90% of the time. If they play in Dodger Stadium, they’re going to have a low offensive fraction 90% of the time. If we were to do that at all, we’d almost have to do a “park-adjusted” offensive fraction. Then you’d find just the opposite: the team that plays in Dodger Stadium would almost always have a high offensive fraction, and the team which plays in Fenway Park would normally have a low offensive fraction.

Besides that, I can just hear myself on a call-in show. “Bill, can you explain to us the concept of a park-adjusted offensive fraction?” Um … well, not really. We’re trying to assemble
simple
information here, clearly understood. Idiot accessible, if you will.

STATS Inc. carries information on the number of starts by pitchers on three days’ rest, which is an interesting thing to know, but I’m afraid that in ten years, we’d have columns of zeroes there.

Let’s deal with some strategies. What the manager does, other than deploy his personnel, is to order strategies—intentional walks, sacrifice bunts, stolen bases, the hit and run. This is the main area that we should address next.

“Intentional walks” already appear on the pitcher’s record, and in general I don’t want to reproduce a lot of information that we already have access to. In this case, because of the importance and clear relevance of the statistic, I’ll make an exception; IBB will be added to the chart. Going back to Tony LaRussa: