Yeah! Who are the great managers anyway?! Well, since we’re going elect 22 MLB managers we’d better find out, huh? As with players, we want to be rigorous and transparent about our process, so I’m going to document some of my thinking right now.
A Whole ‘Nother Ball of Wax
The thing about managers is this: We lack the kind of hard-edged objective information we have for player performance. In fact, statistically, nearly all we really have are top-level indicators of performance:
- Winning percentage
- Pythagorean record
- Pythagorean winning percentage
- World Series appearances
- Playoff appearances
But it turns out this is enough to get started and get us a ways down the path. We can combine these statistics, massage them a couple cool ways, and build a sorting tool that can tell us pretty decently who the best candidates are.
Who Are the Candidates?
But before we put this all together, we need to ask ourselves, what is the field of candidates? Only 694 men have managed in the big leagues. Since we are matching the Hall’s electoral output, we can’t vote on guys who aren’t yet eligible to be considered by their Expansion Era Committee. That means the following 73 are ineligible:
- The 35 men who have managed in 2015
- The 35 managers under 65 years of age in 2014 who managed anytime between 2009 and 2014 (the requirement for the last time this committee met)
- The three managers 65 or older by 2014 who managed in 2013: Davey Johnson, Jim Ley, Charlie Manuel.
That leaves a mere 621 to go. We might as well eliminate manager with fewer than 500 games under their belt. Get rid of them, and we’re down around 210. You know what, let’s knock out everyone under 1,000 games managed and see which important managers go away. Eddie Dyer, Dick Howser, Jimmy Collins, Larry Dierker, Bob Lemon, Yogi, Sam Mele, Joe Altobelli, Johnny Ward, Del Baker, guys like that, really, we’re not going to miss. That gets us down to a nice, reasonable pool of 106 managers.
Oh, but hang on.
Do managers before the 1890s really count? I’m not sure they always do. Maybe some do? Chris Jaffe’s book Evaluating Baseball’s Managers describes how early managers were more like businessmen, traveling secretaries, team treasurers, all rolled together. He says about Jim Mutrie, who owns one of the best managerial records out there: “The most successful business manager was Jim Mutrie, who won two pennants and ended a nine-year stint with a .611 winning percentage. He left most baseball decisions to team captains John Ward and Buck Ewing.” That’s not a manager as we understand them today. On the other hand, Cap Anson, Charlie Comiskey, and Harry Wright were all modern in the sense that they ran the whole show out there. They made the important baseball decisions before, during, and after the game. In fact, they are all well known for strategic innovations in on-field, off-field, and front-office affairs. Even so, it was a very different role, and I’m hesitant to go overboard on those early managers. It’s possible that someone like Wright, the game’s first professional manager, might fit better as a pioneer than a manager. It’s something to watch for as we go along.
Bring Out the Yardsticks
So we can’t measure chemistry and a manager’s ability to rally the troops. But those stats we mentioned above ultimately take them into account. John McGraw could dominate his times because he was better able to keep his players focused and drill them on fundamentals than anyone else of his era. And managers aren’t completely dissimilar to players. Some managers have long careers, some have shorter careers. We need to assess both guys fairly. So here’s how I’ve done it.
First off, I’m going to use the following ten categories, which are broken into five groups.
Regular Season Performance
- Career managerial wins
- Career winning percentage
This is simple, right? In the first case, a guy like Connie Mack gets all the credit for all those many wins. In the second case, a guy like Earl Weaver, who has fewer than half of Mack’s wins, gets to strut his stuff.
- Career variance vs. Pythagenpat
- Career variance vs. Pythagenpat winning percentage
I’m using the more accurate Pythagenpant than the simpler Pythagorean method. Walter Alston’s .542 winning percentage is pretty darned good. But what makes it more impressive is that he finished 62 games above what Pythagenpat would predict. He’s the all-time leader there. Then there’s early Red Sox skipper Bill Carrigan who lasted only a few years but won a couple titles. His actual winning percentage of .494 doesn’t look impressive, but it’s 4% higher than his .477 Pythagenpat winning percentage, tied, with Earl Weaver and Al Lopez, for the biggest percentage difference. In my opinion, some form of evaluation against an expected winning percentage is the crucial fulcrum on which our analysis must sit. Without it, we have facts but no context for them. Jimy Williams’ .535 winning percentage looks pretty nice. Then you find out that he was -24 wins against Pythagenpat and that his Pythagenpat winning percentage was .549, nearly 15 points higher than his actual percentage. Ouch.
The Big Prizes
- Championships won
- Championships won vs expectation
- Finals appearances
- Finals appearances vs expectation
- Playoff appearances
- Playoff appearances vs expectation
Isn’t the Commissioner’s Trophy what every team is really after? We want to count the ringzzzzz. So that’s what we’re looking at here, and we want to count the important steps along the way. But not as much. I’m weighting championships fully, World Series/Finals appearances half as much as winning it, and making the playoffs half as much yet. Now, you might ask, what do I mean by opportunity? I’m using pretty much the same method as I did in my article about which cities had suffered or been spoiled. The idea is simple: assuming an even distribution to all teams over time, find the probability of winning a World Series/League Championship/Playoff Spot each year based on the number of teams in the league and spots up for grabs. (Note: I only counted years where a guy managed 100 or more games.) Joe McCarthy won 7 titles. Based on the 16-team leagues he managed in with the winners facing each other in October, Marse Joe had a probability of winning 1.5 titles in the seasons he managed. That’s a difference of 5.5 titles. Wow. On the other hand, Gene Mauch, Wilbert Robinson, and Jimmie Dykes never won a title during very long careers and ended up each with a net of -1.1 titles versus expectation.
OK, so that’s what we’re looking at, and we obviously will want to compare each of these guys to the others, so we’re going to evaluate against the average of the pool. But these categories will have a heterogeneous range of values, especially since some are percentages and others raw totals. Looks like a job for z-scores. I’m not a statistician, but what z-scores do is provided some standardization among heterogeneous data.
In each category, we will:
A) Find the pool’s average
B) Subtract that average from the manager’s value in that category
C) Find the standard deviation in that category
D) Divide B by C to get the z-score.
E) Add up all the z-scores for the manager’s score.
If you wanted to know, here’s the averages for the group:
- Career managerial wins: 1094
- Career winning percentage: .512
- Career variance vs Pythagenpat: +9.6
- Career variance vs Pythagenpat percentage: 0.999
- Championships won: 1.1
- Championships vs expectations: +0.3
- Finals appearances: 1.7
- Finals appearances vs expectations: +0.4
- Playoff appearances: 1.3
- Playoff appearances vs expectations: +0.7
This works out pretty good, but I want to go one better.
I want to push the best managers up the list as quickly as possible. So I’m going to add a bonus system. If you finish in the top 22 of each category (to match the 22 managers we seek), you get a bonus.
That bonus is figured this way
A) If a manager ranks 23rd or lower in a category, no bonus
B) If a manager ranks 22nd or better in a category, then
(23 – rank in category) / 22.
Of course, this means that if you rank first, you get 1 bonus point. If you rank 22nd, you get about .045. Add the bonuses to step E above, and you’ve got your final managerial score.
This is a sifting tool, we want to know quickly who the guys to pay the most attention to are. Here’s everyone who finished with a manager score above average in this group, an asterisk (*) denotes that they are in the Hall of Fame in Cooperstown as a manager. A plus (+) denotes they are enshrined as a pioneer or executive but that their plaque makes prominent mention of managerial contributions. Remember that this doesn’t mean that this is the order I would necessarily rank them. There’s always, always context to consider.
1. Joe McCarthy*: 24.7
2. Joe Toree*: 19.0
3. John McGraw*: 18.3
4. Tony La Russa*: 17.5
5. Casey Stengel*: 17.4
6. Walter Alston*: 17.3
7. Sparky Anderson*: 15.5
8. Connie Mack*: 13.8
9. Bobby Cox*: 13.6
10. Frank Selee*: 13.5
11. Earl Weaver*: 12.5
12. Miller Huggins*: 12.0
13. Harry Wright+: 10.3
14. Frank Chance*: 9.8
15. Billy Southworth*: 9.3
16. Cap Anson: 8.2
17. Charlie Comiskey+: 7.7
18. Jim Mutrie: 7.6
19. Bill McKechnie*: 7.4
20. Fred Clark: 7.4
21. Tommy Lasorda*: 7.3
22. Dick Williams*: 6.8
23. Al Lopez*: 6.2
24. Whitey Herzog*: 6.2
25. Billy Martin: 6.1
26. Ned Hanlon*: 5.6
27. Ralph Houk: 4.9
28. Danny Murtaugh: 4.8
29. Bill Carrigan: 3.7
30. Cito Gaston: 3.7
31. Pat Moran: 3.4
32. Hughie Jennings*: 3.4
33. Lou Piniella: 3.3
34. Leo Durocher*: 3.1
35. Bill Terry: 2.9
36. Steve O’Neill: 2.3
37. Tom Kelly: 1.5
38. Bucky Harris*: 1.1
39. Charlie Grimm: 1.1
40. Clark Griffith+: 0.7
41. Mike Hargove: 0.7
42. Red Schoendienst: 0.5
43. Jack McKeon: 0.3
44. Al Dark: 0.4
45. Joe Cronin: 0.3
46. Herman Franks: 0.2
The astute among you see that one Hall of Fame manager is not above average for this group:
53. Wilbert Robinson: -1.0
Yup, an otherwise utterly average manager except that he failed to win a World Series as mentioned above.
At the other end of things from the amazing Joe McCarthy (seemingly the Babe Ruth of managers) are these amazingly bad pilots:
102. Hugh Duffy: -7.9
103. Buddy Bell: -9.4
104. Jimmie Wilson: -10.6
These guys stand out for their utter awfulness. Not one of them scores positively in any of our categories (and, obviously, earns no bonus points). They managed in parts of 26 seasons among them with a combined record of 1547–2130, or 583 games below .500. Their combined .421 winning percentage as a trio means that their typical team was equivalent to a 68-game-winner under the modern schedule. None made the post-season, and the best finish among them was 4th in an eight-team league or third in a five-team division. These three finished a combined -116 games below Pythagenpat with winning percentages, in the amalgam, 7% lower than their expected winning percentage. Remember that the best managers only got as high as 4% higher than Pythagenpat. I think this stuff speaks for itself.
By the way, if you are wondering about how some not-yet-eligible managers stack up, here’s a no-brainer: Put your money on Bruce Bochy. And here’s a tout: Watch out for Terry Francona. Yeah, I know, you didn’t really need me to tell you that, did you?
What does this list tell me? That we’ve probably done a good job of modeling what a great managerial career looks like. Not that this is too hard, nor even too different from what the Hall’s Veterans Committees have done more intuitively. But it also tells me that we have some interesting decisions ahead of us. Are Comiskey, Anson, and Wright all really great managers? Or is there something about their times that we may need to watch out for? Are Hall of Famers Durocher and Jennings not as great as we thought? And, gee, what about Bucky Harris? We’ve got some great discussions coming up!