Article: Top Ten Overachieving Managers: All-Star Selection Bias

[Teflon]

On a previous lunch hour, between bites of an All-P sandwich* I was wondering if there was a usable correlation between the number of All-Stars teams have and an expected number of wins. I thought if a team had a lot of All-Stars and lower than expected wins, this might suggest lackluster managing. Higher than average expected wins per All-Star, perhaps good managing.It turns out there actually is a very strong step up in average winning percentage for each incremental All-Star a team has. Roughly this: one All-Star – average winning percent of .458; two All-Stars, .501; three All-Stars, .542; four All-Stars, .557, and so on.

The problem with using this information is the incredible market bias in All-Star selections. To be able to evaluate Yankees managers and Twins managers on common ground, each team’s positive or negative bias in All-Star selections has to be compensated for.

 

To do this, I counted the number of All-Star selections each year and totaled the number of league wins. I then computed each year’s average wins per All-Star and applied this to each team’s winning percentage to determine the equitable number of All-Stars each team should have had. Summing the actual number of All-Stars each team has had through its history and dividing this by the sum of equitable All-Stars creates an index value per team than can be used for the aforementioned manager evaluation.

 

Which team has the biggest bias in All-Star team selection? Do I really have to ask? Of course, it’s the Yankees with nearly 52% more All-Stars than a fair distribution suggests, given the team’s success. This is far and away the highest bias, as the closest team, the Cardinals, are at +23%. Rounding out the top 5 are the Red Sox at +21%, and the Dodgers and the Reds both at 13%. Conversely, the Rodney Dangerfield teams of All-Star selections are led by the Rays at -20%, the A’s and Astros at -19%, the Pirates at -18%, and the Royals at -17%. The Twins, if you’re curious, are at -15%, which is 9th worst.

 

If you’re curious about which managers overachieved or underachieved in their careers relative to the number of All-Stars their teams had, adjusted for All-Star bias, here are looks at the top 10 and bottom 10 calculated during today’s lunch hour between bites of a Waldorf salad. Results are based on average seasonal difference between actual wins and expected wins. And the manager had to be the manager for the majority of a team’s games to be credited for that season and had to manage at least 5 seasons.

 

The Good:

1. Joe McCarthy +12.7 wins per season.

2. Billy Southworth +11.2 wins

3. Joe Girardi +9.9 wins

4. Bobby Cox +7.8 wins

5. Al Lopez +7.7 wins

6. Earl Weaver +7.7 wins

7. Walter Alston + 7.7 wins

8. Pete Rose + 6.7 wins

9. Billy Martin + 6.6 wins

10. Jimy Williams +6.5 wins

 

The Not So Good:

1. Connie Mack -12.2 wins

2. Jimmie Wilson -11.8 wins

3. Billy Meyer -9.7 wins

4. Preston Gomez -9.7 wins

5. Buddy Bell -9.3 wins

6. Ossie Bluege -8.1 wins

7. Fred Haney -7.7 wins

8. Del Crandall -6.5 wins

9. Manny Acta -6.1 wins

10. Mary Marion -6.0 wins

 

Twins managers:

1. Billy Martin +6.7 wins

2. Frank Quilici +3.2 wins

3. Ron Gardenhire +1.5 wins

4. Gene Mauch +0.7 wins

5. Ray Miller -0.3 wins

6. Sam Mele -1.2 wins

7. Tom Kelly -2.4 wins

8. Cal Ermer -2.6 wins

9. Billy Gardner -3.9 wins

10. Bill Rigney -5.4 wins

* pastrami, prosciutto, provolone, peppers, poupon and pickles on pumpernickel

 

Click here to view the article

[hybridbear]

Fascinating info! Thanks!

[old nurse]

Every team gets an All Star, even if the person is not deserving. Fans select players, even if they are not deserving. Host team bias comes in. Ii is a unique way to look at things.

[Kirby_waved_at_me]

That sandwich sounds amazing - do you serve it hot or cold?

 

Also, something about baseball.

[Cris E]

The bias isn't in the voting for starting lineups. There are only nine or ten names there out of dozens. The serious bias is in the selection of pitching and backups. The best teams with the most wins get the most all-stars, but the causality might not be players bringing wins as much as wins making all-stars.

 

There are two forces that multiply each other to concentrate the player selections on the best teams: the manager comes from the team that won the World Series, and the manager gets to select his guys for many secondary roles.  If every team got to provide the manager at some point (similar to site selection or One Player Per Team) then this might start to even out. But team selection stays in the hands of the winning teams and they'll keep appointing their own guys as much as possible. 

 

It turns out all-stars are usually not flash in the pan meteors, so having a bunch of them is also predictive in how many games you might win next year. But the the non-stats factor in how they're selected that makes a straight wins formula hard.  Second place is waayyy behind first place.

 

TRDR version:  Look through the over-acheivers and count how many WS wins there are versus the other lists.

Edited August 25, 2014 by Cris E

[notoriousgod71]

Does this take into effect players selected for all-star games that did not participate (such as any pitcher that pitches on Sunday)? Injury replacements? How did you factor in increased all-star game rosters?

[Paul Pleiss]

Does this sandwich have room for a sponsor? Paul's pastrami, prosciutto, provolone, peppers, poupon and pickles on pumpernickel sounds good to me.

[Kirby_waved_at_me]

Preach, Paul.