THE BOOK--Playing The Percentages In Baseball

That is a pretty reasonable article.  None of the usual stupid things you find in sports articles in mainstream non (for the most part)-sports newspapers and magazines.

Off the top of my head, I would think that if a manager and his coaching staff are going to have a large influence on performance, it will be with the pitchers and not with the position players.  And I would also think that the responsibility falls primarily with the pitching coach and not with the manager.

I don’t think, however, that because teams pay managers within the range of one marginal win, that’s what they think managers are necessarily worth (or that is what they ARE worth).  There are many reasons for and influences on the various pay scales and ranges in sports, and the marginal win value of the personnel is only one of them.

Good point.  Mostly, it is supply/demand.  There is an endless stream of people wanting to be managers.  And, there is a limited way to evaluate a manager.

Even if a manager should be paid 4.4MM per actual win, the ability of management to decipher how much a manager adds is very limited. 

Terry Francona could ask for 9MM, and the Redsox could show that he adds 2 wins a year, but they are not certain of that, and will likely regress that to 1 win or less.

But even if there is a strem of thousands of managers applying for the job, the salaries will be determined by the 30th best.  Anyone one win better than that 30th guy should make 4.4 million more than him. 

I agree that it would be hard to show that a manager contributes even as much as a win a year.  And, as Bill James has said, different teams call for different managerial skills, so you’re looking for someone who fits in with the team’s needs.  (With a player, runs are runs.) So that makes it even harder to figure the manager’s worth.

Two questions:

What do you guys think about how much a manager benefits his team?  Do you think there’s any manager out there (that you can identify) who’s worth a lot?  Serious question, I don’t know the answer.

And anyone know where to find Gassko’s original article?  I did a google search and came up empty.  I’m especially interested in the “getting the most out of his players” aspect.

Oh, wait, there’s a Gassko article in the Hardball Times 2008 ... I bet that’s the one.

Yep, Phil, it is.

There is some serious risk of “double-counting” things here. Anything that the manager does that makes a difference in the players measurable stats will almost by default be atributed to the player(s) and not to the manager.

I’m convinced that (for example) the players OPS is affected somewhat by their mental/emotional state which is certanly not independent of what the manager does. But to quantify this and separate it from the random noise and other effects that impacts performance seems almost impossible.

I think David’s method of measuring managers’ impact on player performance is probably an improvement over Jaffe’s, and certainly agree with his decision to ignore both wins-over-pythag and “extra” runs created (over context-neutral performance).  However, I think there are three problems worth addressing in a version 2.0 of the analysis.

First, the age curves create a bias in favor of long-career players and against short-career players.  A player who only plays in MLB from, say, age 25 to 30, is guaranteed to have a negative total runs-vs.-self.  The age curve requires him to outperform his own career average in those years, which is of course impossible.  On the flip side, star players with long careers will tend to post great runs-vs-self totals in their prime years, as their career average is lowered by many post-peak years of play (but the age curve—flattened by inclusion of so many non-stars—won’t discount the peak performance enough).  So, managers who get a lot of playing time from star players in their prime have an inherent advantage. (I think David and I had an exchange on this issue before, but not sure...)

Second, managers with good defenses will look artificially good, as their pitchers’ wOBA-vs-self will tend to be low.  I would guess that Weaver and Cox benefit from this bias by a non-trivial amount.  Torre, perhaps, suffers as a result.

Third (and least important, I’m sure), managers with a disproportionate number of IP from relief pitchers who used to start will have an edge, due to the “reliever advantage.”

Note also that at least the first two factors will tend to reward winning managers (good defense and great players in their prime both = more wins).  So this will tend to make the results feel intuitively correct.