In the comments of Part I, a reader suggested using a measure of team talent other than actual wins in the calculation of payroll efficiency.  Not a bad idea.  I'm going to use Baseball Prospectus' third-order wins metric, which adjusts for run differential (via EqA) and strength of schedule.  The average third-order win totals for NL and AL teams were 78.5 and 84.8, respectively.

The merits of third-order wins and other Pythagorean-like win estimators have been debated before, so I won't mention much about them here.  There are pluses and minuses, but it's all a gray area.  From reading a lot of things written on the topic and crunching a lot of numbers throughout the years, I tend to think that, on average, 40% of the difference between actual record and Pythag record is due to skill (extreme bullpen leverage, non-SB baserunning, and doing the little things) while 60% is luck (clutch hitting, the way batting events combine into runs, run distribution, alignment of RS and RA, etc.)  Certainly it can be (mostly) all luck or all skill for a given team in a given year.

Anyway, the table below uses the same methodology as Part II.  Using the logistic model, I've calculated the expected payroll to produce a team with that much talent and compared it to each team's actual payroll in order to find their savings or wastings.  All dollars are in millions:

Notes:

• No suprise the Rays look good again, although their actual win total was a bit above their true talent level.
• The Red Sox vault up to second in efficiency.  Going into 2009, there's no doubt the Sox should be the favorite in the AL East, if you look beyond 2008's actual record.
• In general, teams are more clustered in the middle.  That makes sense, because third-order wins attempt to remove luck, and luck causes the distribution to spread wider.  It adds to the overall variance.

We can also compare the savings computed using third-order wins to the savings computed using actual wins.  This is the same thing as comparing the two win totals directly, but puts it on the payroll scale.  A positive difference means the third-order method gives the team more savings (or less wastings):

Notes:

• The top of this list reads like a list of teams you shouldn't forget about going into next season.  And the bottom is a list of teams that will likely disappoint.
• The Angels and Twins do have a habit of out-performing their Pythag, mostly due to team baserunning skills.  I still expect a fair amount of regression, just not all the way.  And don't expect their clutch hitting skills to continue.
• The fall of Ed Wade will not be pretty.