Bill James Guide to Baseball Managers, The (22 page)

The team actually went 84–76, 21 games better than expected. This is very rare. Only a few managers will ever have a team finish 20 games ahead of where they could have been expected to finish. In 1996, two National League teams finished 10 games better than expected: San Diego at + 15, and St. Louis at +12.

We figured win expectations for each season of the careers of a large number of managers. The resulting records are given in the chart (below):

I should stop and explain what all of that means, in case any of it is unclear. Felipe Alou’s best season, by this method, was 1994, when Montreal was 74–40, the best record in baseball at the time the season was stopped by the strike. This was 12 games better than the team could have been expected to play in 114 games, based on their performance in the previous seasons.

Alou has had four good seasons so far in his career, and only one bad season, with a good season being any season in which the manager’s team outperforms expectations by three games or more, and a bad season being any season in which the team fails by three games or more to meet their expectations. From plus two to minus two, the season is neither good nor bad.

The next three columns are just the manager’s career won-lost record and winning percentage, and the final column is his career plus or minus versus expectations. Alou’s Expos were +11 in 1992, +7 in 1993, +12 in 1994, and +6 in 1996, but -17 in 1995, for a total of +19 games.

That’s the beginning of the alphabetical chart; I’ll switch now to putting managers in order of accomplishment.

I have about fifty notes of things I was supposed to explain about this chart; I’ll try to keep them under control.

1) We didn’t figure the records for the true nineteenth-century managers, for reasons you can figure out if you care.

2) This evaluation is careful, logical, and has a scientific basis, since the win expectation can be objectively verified by anyone who chooses to study the issue. Two teams win 63 games, 67 games, 66 games each, then hire new managers. One of those teams then wins 85 games, and the other wins 55. It is reasonable to conclude that one manager has had a better year than the other. This method is just a way of systematizing those type of comparisons. We can’t compare each manager’s season to a large field of teams coming off a string of identical seasons, of course, because there aren’t that many identical strings of seasons. By using this method, we can establish what the field of teams coming off a string of identical seasons
would
do, if there were such a field.

3) At the same time, let’s not overstate the value of the method. There could be a million reasons why a team could have a bad year or a good year, other than what the manager has done.

Walter Alston’s teams were 427 games over .500 in his career, yet Walter scores at only +26. Why? Because his teams were always expected to do well. When his teams played well one year, they were expected to play well the next, so Walter got little credit for it. This may not be entirely fair to managers who stay in one job for a long period of time, as opposed to those who bounce from job to job, like Dick Williams (+92) and Billy Martin (+98).

4) In calculating expected wins, expansion years are a pain in the rear. For an expansion team, we just entered the “last year’s record” at 54–108, and the previous year’s records as the same. This creates an expected first-year winning percentage of .375, and the actual first-year winning percentage of all expansion teams is like .376 or something, so that’s fine.

The problem is, it throws the league off-center. The “expected losses” for a league will outweigh the expected wins, by a large enough margin that it makes the data look funny. To prevent that from happening, you have to adjust upward the expected wins for all of the established teams, not only for the expansion year, but for the subsequent season, as well.

This is a petty detail, and I realize that 99% of you don’t care, but the assumption of verifiable research is that someone else could duplicate your work, if they chose to, and reach the same conclusions. To make that viable, I have to explain the petty details.

5) The biggest surprise in the data is Bobby Cox. Bobby Cox, evaluated by comparing the performance of his teams to the expected performance of his teams, is closing in on the distinction of being the greatest manager of all time.

Will he make it? I wouldn’t bet on it. It’s a zero-sum game; an average team in any season is at zero. Cox is as likely to move
down
the list as he is to move up. Sparky Anderson, at one time, was + 70. Cox, because of Atlanta’s performance in recent years, has very high expectations to deal with. In 1996, Atlanta won 96 games. That’s +0; they were expected to win 96.

Nonetheless, it’s an impressive record. Cox has had nine good seasons as a manager, and he’s never
really
had a bad one. His worst season as a manager was 1979, when he went 66–94 with a team that could have been expected to win 70. That qualifies as a bad season in the chart, but barely.

What makes Bobby win? The most obvious thing is that he has acquired Earl Weaver’s uncanny ability to keep his starting pitchers healthy and productive. When Cox took over the Braves in 1978, they had posted the worst earned run average in major league baseball, 4.85. They cut that to 4.08 in 1978, to 3.77 in 1980, and 3.45 in 1981, better than the National League average.

He joined Toronto in 1982; they had been eleventh in the league in ERA. Their team ERA improved, under Cox, to fifth in the league in 1982, to third in 1984, and to first in 1985, his final year with the team. He inherited a team with three starting pitchers, Dave Stieb, Luis Leal, and Jim Clancy. Leal got hurt, but four years later Stieb and Clancy were still in the rotation, and both pitching better than they had when Cox arrived. They had been joined by Doyle Alexander and Jimmy Key, giving Cox the best starting rotation in the American League at that time.

In his second term in Atlanta, of course, he has done the same act, only better. John Smoltz, Tom Glavine, and Steve Avery were there when Cox returned in 1990. Five years later, they were still there, pitching better than they had in 1990, while the addition of Greg Maddux had created the most formidable starting rotation in many years.

6) Bobby Cox was +8 his first time in Atlanta, +49 in Toronto, and is +58 since returning to Atlanta (1990–1996).

Joe McCarthy was +36 in Chicago, +74 in New York, and +16 in Boston.

Casey Stengel was +55 with the Yankees, but -8 in Brooklyn, -25 in Boston, and -31 with the Mets.

Connie Mack was +1 with Pittsburgh in the 1890s, +69 from 1901 to 1914, -102 from 1915 to 1921, +92 from 1922 to 1931, and -109 from 1932 to the end of his career.

7) This book is not about how managers rate, and to the extent that I would rate managers, I would use the point system introduced earlier in this article, rather than the plus/minus system.

Nonetheless, I do believe there is some validity to the method. Let’s look, for example, at the managers who had losing records in their careers, but who score on the plus side in this system.

That would include Bucky Harris (61 games under .500, but 52 games better than expectation), Gil Hodges (93 games under .500, but 28 games better than could have been expected), Gene Mauch (135 games under .500, but 18 games better than expected), plus George Bamberger, Bobby Valentine, Luke Sewell, Bill Carrigan, Harry Craft, Burt Shotton, and several others.

I believe, in all cases except one (Fred Haney) that it would be fair to say that these men were successful managers, but just didn’t have a lot to work with for most of their careers—thus, I believe that what the method is telling us is essentially correct.

On the other hand, we have managers who were over .500, but under expectation—Mayo Smith, Eddie Stanky, Red Schoendienst. In general, I would have to agree that an accurate assessment of these men’s records is that they won, but given the teams that they had, they didn’t win as often as might have been expected. It’s a subjective judgment, but in general, I’ll buy it.

The 1930s was the last decade in which player/managers were common; between 1935 and 1955 they made a strong move toward extinction.

You may have heard it said that the player/managers were rarely effective after they stopped playing, and became just plain managers. Lou Boudreau, for example, managed three teams after he stopped playing, which was an unpleasant experience for all concerned. Frank Chance and Fred Clarke, the great player/managers of the first decade of the twentieth century, both developed a habit of getting the hell beat out of them as soon as they stopped playing, as did Bucky Harris and Rogers Hornsby and Frankie Frisch.

I have concluded that this belief is essentially false, for three reasons.

1) It’s inherently inexplicable,

2) The statement attributes to one class of managers what is in fact generally true of
all
managers, and

3) The statement relies on a “false sorting” which makes the pattern seem stronger than it is.

First, why would player/managers become unable to manage after they became unable to play? You can make a
general
statement about player/managers having to adjust their approach to managing from the dugout, but when you try to make it specific, you realize it doesn’t make any sense. What, specifically, are these managers going to forget how to do? Signal a bunt?

Second,
all
managers, as a group, are most effective in their early years on the job.

I did a study of 103 managers who managed at least 600 major league games, a group basically including all twentieth-century managers who had significant careers and are now retired. The study documented something which is apparent if you just look at the records. A huge percentage of managers have their best seasons

a) when they first get a chance to manage, and

b) in their first years on a new job.

This is almost a universal rule. There might be ten exceptions in the century—Casey Stengel, for example, and Gene Mauch. But the teams managed by these 103 managers outperformed expectations (see “
Ranking Managers
”) by 221 games in their first seasons as major league managers (2.15 games per manager), and by a whopping 494 games (4.80 games per manager) in their second seasons, which in many cases was the first full season.

They continued to be enormously effective in their third seasons (+3.16 games, on average), and fourth (+2.37 games)—but after that, they went into idle. In seasons five through ten, these 103 managers improved their teams, on average, but by less than one game. In seasons eleven through fifteen, the teams managed by these men did not outperform expectations at all, and after year sixteen, they tended slightly to
under
-perform the natural expectation of the team.

I’ll return to this theme in a moment, but what I am saying here is that since
all
managers have a strong tendency to lose effectiveness as they continue in the job, the fact that
player/managers
have the same tendency is not noteworthy.

And third, think about Miller Huggins. You may not even know that Miller Huggins began managing while he was still playing, and in fact was quite a successful player/manager. You may not know this, however, because everyone’s image of Miller Huggins is formed by his most successful seasons, which were his years with the Yankees in the 1920s.

There are other examples like this, such as Leo Durocher. If a manager begins as a player/manager and has his best years as a player/manager, we think of him as a player/manager, like Lou Boudreau. But if he has his best years as a dugout manager, we classify him as a dugout manager. This subconscious grouping tends to sharply reduce the post-career managerial success of player/managers, because if they
are
good, we tend not to think of them as player/managers at all.

The most interesting question here is why managers lose effectiveness as they remain in their job. I find myself discussing this constantly on talk shows, whenever a team changes managers. Some fan will always call, alleging that the fired manager was a scapegoat for the problems of the team. Mr. Wombat was a good manager two or three years ago, wasn’t he? He won the league manager-of-the year award just two years ago, didn’t he? If he was a good manager then, why isn’t he a good manager now?

Nonetheless, the most obvious fact about managers is that almost all managers become ineffective after two or three years in a position. I will cite a few examples of this, but I have misgivings about doing so, because citing examples tends to obscure the fact that
almost all
managers are “examples” of the same phenomenon:

Joe Altobelli
improved the San Francisco Giants by 12 games his second year there, 1978. By the middle of the next summer, his clubhouse was in chaos.

He was hired to manage Baltimore in 1983 and won the World Series in his first season there. He was fired a year and a half later. The only seasons of his career in which his teams exceeded expectations were 1978 and 1983.

George Bamberger
got the Milwaukee Brewers in 1978 and was phenomenally effective for two years, with his team exceeding expected wins by 21 games in 1978, and 12 more in 1979. He managed for five more seasons or parts of seasons after that, but all five of those teams failed to meet expectations.

Roger Craig
took over San Diego in 1978 and had a great first year. The team collapsed in 1979.

Craig assumed command of the San Francisco Giants in late 1985 and the team exceeded expectations by 14 games in 1986, and 12 more in 1987. They were on target in 1988, and +9 in 1989, Craig’s fourth year there. But the team failed to meet expectations in 1990, 1991, and again in 1992.

Jim Frey
got his long-awaited opportunity to manage in Kansas City in 1980. He took the 1980 Royals to the World Series. By August of 1981, he was staggering around in a daze.

He managed the Chicago Cubs in 1984, and the team exceeded expectations by 24 games, moving from fifth to first in one season. But he managed them for two more years after that, during which they returned to the mediocrity from which they had sprung.

Clark Griffith
managed the White Sox, Highlanders, Reds, and Senators from 1901 to 1920. All four of these teams improved dramatically in Griffith’s first season as their manager, and two of them continued to improve in his second season. But all of them regressed after that.

The Senators are the most dramatic example. The Senators went 42–110 in 1909, 66–85 in 1910, and 64–90 in 1911. On this basis, they could have been expected to go 65–89 in 1912, the season in which Griffith arrived.

Under Griffith’s management, they went 91–61, finishing second in the American League. They won 90 games again in 1913, and stayed over .500 in 1914 and 1915. They dropped under .500 in 1916 and 1917, and by 1919, they were back in seventh place.

Pinky Higgins’s
Boston Red Sox were +10 in 1955, his first season as a manager; they were expected to win 74, but won 84. The Red Sox remained in the plus category, by smaller margins, in 1956 and 1957—but Higgins managed them for parts of five more seasons, during all of which the Red Sox failed to win as many games as expected.

Bob Lemon
was +17 his first full year in Kansas City, and +15 his first year in Chicago. He was +9 his first year in New York, when he took over in late July. The rest of his career, his teams consistently failed to meet their expectations.

Billy Martin,
of course, improved every team he ever managed in his first season in control, usually by huge amounts. Within a year or two, all of those teams were ready to get rid of him.

Chuck Tanner,
after two years on the job, was +28, his 1971–1972 White Sox having exceeded expectations by 13 and 15 games. He wound up his career -28.

Bobby Valentine
improved the Rangers by 17 games in his first full season as their manager, 1986. He managed them until 1992, and they never had another season as good.

Ted Williams
impressed the baseball world as a rookie manager in 1969, when he took the Senators to a record of 86–76. He managed them three more seasons, with records of 70–92, 63–96, and 54–100.

So managers have a strong tendency to lose effectiveness over time. The question now is, why?

I have probably thought more about that issue, over the years, than about any other issue discussed in this book. I have concluded that there are many reasons why this occurs, which unfortunately interact in ways that make them impossible to classify.

If a manager is successful, he changes the needs of the organization. By so doing, he often makes himself obsolete
.

The most important question that a manager asks is “What needs to be changed around here?” Any manager, over time, loses the ability to see what needs to be changed.

Managers, and ball clubs, can be split A/B along many different axes—for example, high pressure/low pressure, platoon/nonplatoon, aggressive/patient, young/experienced. All of those categories are limiting. New managers are effective because they tend to pull their teams out of the rut created by overemphasizing any one category.

For example, a high-pressure manager will almost always, in time, find himself with an emotionally charged, uptight clubhouse in which no one feels at home. The more high-pressure a manager is, the more quickly that point will be reached. When that manager is replaced by a relaxed, low-pressure manager, almost everyone on the team will play better for some period of time.

Conversely, when a low-pressure manager remains in charge of a team for a period of time, the team will almost always lose focus. The players won’t work as hard, and they won’t concentrate on the things that the team needs them to work on. They’ll lose their fear, and when they lose their fear, they’ll lose their edge. When a high-pressure manager takes over, everyone plays harder, and the team plays better.

To sustain success over a period of time, a manager must avoid both of these traps: He must be neither too low-pressure nor too high-pressure. This is nearly impossible. Joe McCarthy was able to do it; Bobby Cox has been able to do it.

Thus, teams over time are whipsawed between high-pressure and low-pressure managers. The New York Yankees in the years of Billy Martin and Bob Lemon are the classic example, Martin being the highest of high-pressure managers, and Lemon the most casual of low-pressure managers. Each was effective because he replaced the other.

Similar things happen on all of the other axes. If a team
doesn’t
use its bench, the bench will atrophy, and the team will become vulnerable to injuries. If a team overuses its bench, they’ll have difficultly developing stars. The manager, to remain effective, must find exactly the right balance.

If the team uses too many veterans, within a couple of years age will catch up with them. In that case, they probably won’t begin rebuilding until they go to another manager.

But if the team uses too many young players, they may have difficulty getting over the hump. In that case, they might not “arrive,” as a team, until another manager comes along who brings in some veteran leadership.

This is not to say that all managerial styles are equal. I believe, rightly or wrongly, that low pressure is better than high pressure, young is better than old, and using the bench is better than not using the bench. My point is that there are many such orientations, and each of them creates pitfalls for the manager.

Even apart from those pitfalls, there are profound reasons why new managers tend to be more effective than established managers. There is the manager’s loyalty to his players. A new manager owes nobody nothing. He can bench or release unproductive players without apology. An established manager can’t do that—not only because of his own reluctance to break faith with players who have given him their best efforts, but because of what it means to the rest of the team. Good teams have the attitude that “we are all in this together.” When a manager who has been a part of that “We” suddenly decides to toss Charlie out of the boat, it shakes the group identity of those who remain.

If a
new
manager throws Charlie out of the boat, everybody else says, “Uh oh. My job isn’t safe here; I’d better bust my hump trying to impress this guy.” If an
old
manager throws Charlie out of the boat, it’s a betrayal. Everybody else says, “You can’t trust that guy anymore. Charlie busted his butt for him, and look what happened to him.”

Or ask this, “Why are first impressions lasting?” Because that’s when you’re paying the most attention. That fact gives a new manager a tremendous edge in trying to send a message to his team.

Even that is not the end of his advantage.

In the United States Army, the first sergeant is, in essence, the company’s “manager.” When I was in the army, it happened one time that I was assigned to one company for a longer period than the normal thirteen-month rotation, and for this reason I was with the company when a new first sergeant came in, and still there a year later when the next new first sergeant arrived.

They were very different men, but both of them did many of the same type of things, which were intended to convey to the company this simple message: We care, and we’re trying to make things better around here. One first sergeant, to demonstrate that he cared, took down some old, tawdry curtains which hung in the NCO club. The next first sergeant, to demonstrate the same point, put the same curtains back up.

One first sergeant, to show that he was interested in improving things, changed the seating in the mess hall. The next first sergeant shifted it again—more or less back to what it was.

One first sergeant shifted the exercise routine from the morning to the evening. The next first sergeant moved it back to the morning.

And both times, it worked. The attitude of the company, which was very negative as a starting point, improved tremendously when the first first sergeant took down the curtains in the NCO club, changed the seating in the mess hall, and moved the exercise period to 4:00. Over the course of the year the company grew stale, petty problems ground us down, and the attitude turned sour. When the other new first sergeant came in, he did exactly the opposite things—but had exactly the same effect.