RealClearSports
Advertisement

MLB Insights


August 11, 2009 10:22 PM

The General Manager Study - Part Two

Last time I introduced you to the different categories of GM's I had created, and which GM's fell into each category, along with their career winning percentage.  Now it is time to take a look at some of the deeper numbers. 

I ran regressions on every possible variable, but could not find any correlation.  Season by season winning percentage, careers wins, career winning percentage, the list goes on; nothing gave me a number worth writing about.  This is not surprising.  Even guys who are considered top GM's like Billy Beane or Theo Epstein could not turn the Washington Nationals around overnight, no matter which category they are in.  There are also good GM's and bad GM's (in terms of winning percentage) in each category, making it difficult to pick up any type of correlation.

However, this does not mean we cannot have a little fun with the numbers.  Here are the rankings of each category by career winning percentage:

Stats - .525

Law - .503

Baseball Operations - .502 

Baseball Other - .499

Player - .498

Scout - .487

Coach - .484

As a whole, the statistic oriented GM's have fared best, with a .023 higher percentage than the closest competitor, Baseball Operations.  I have discounted Law because it consists of only one GM and does not give enough data to be significant.  Obviously, these numbers must be taken with a grain of salt.  A team that hires a statistical GM is by no means guaranteed a better than .500 record.  Only 23 of the 36 seasons governed by a statistical GM have resulted in such a record, good for a 64% clip.  This was all but proven by the lack of correlation in the numbers that I discussed earlier.  There is no formula that can even begin to predict a teams winning percentage based on the type of GM they hire.

Now let's take a look a little deeper inside each category.

Stats

Average Record: 85-77

Wins Standard Deviation: 10.5.  This means that 68 % of the time, assuming normal distribution, they would win between 75-95 games.

Scouts

Average Record: 79-83

Wins Standard Deviation: 11.6.  So between 67-91 wins

Player

Average Record: 81-81

Wins Standard Deviation: 9.8, which means between 71-91 wins

Coach

Average Record: 78-84

Wins Standard Deviation: 13.0, or between 65 and 91 wins 68% of the time

Law

This category is not even worth discussing.  There were so few data points that the standard deviation is ridiculous and it would be silly to try and draw any inferences.

Baseball Operations

Average Record: 81-81

Wins Standard Deviation: 13.0, good for between 68-94 wins.

Baseball Other

Average Record: 81-81

Wins Standard Deviation: 8.8, between 72-90 wins.

At first glance this might not seem like much, but I think there is something that can be pulled from this.  Three of the categories max out at 91 wins one deviation above the mean, and a fourth at 90, while stats is tops at 95 wins.  Here is how that many wins would translate into playoff appearances over the course of the study if I inserted four teams with that many wins into each league, each year.  Remember, this is only when they finish more than one standard deviation above the mean, which happens only about 16% of the time.

95 wins

Playoffs: 40

Tiebreaker: 13

No Playoffs: 35   

91 Wins

Playoffs: 21

Tiebreaker: 4

No Playoffs: 63

90 Wins

Playoffs: 14

Tiebreaker: 5

No Playoffs: 69

As you can see, those extra four wins more than double the chances of making the playoffs. As many of you commented in my Ricciardi article, there is a lot more to success than regular season wins.  As we can see here, the statistical GM's were expected to be much more successful in terms of making the post-season.  While this cannot be used as a predictor of future success, I think it is interesting nonetheless.

That is all I will leave you with today.  I am a little disappointed that I could not find any correlation between GM type and success, so I will try a few other variables, most notable, the quality of the roster left by their predecessor.  I will most likely tackle this by comparing how many wins the team had the season before the new GM took over.  I do not know if I will find anything, but I will report back with the results either way.

A Member Of