| | I'm an NL fan, but the fact that the AL has been stronger for the past several years is inescapable. The AL is ahead in interleague play so far in 2009, and (thank you to Rob Neyer for printing these totals), in the past 4 years, the AL has been:
149-103 (2008)
137-115 (2007)
154-98 (2006)
136-116 (2005)
in interleague play. In terms of winning percentages, that's .591, .543, .611, and .540, for an overall total of .571. This is very, very statistically significant. The AL is stronger, deal with it.
The more interesting question, though, is WHY the AL is stronger. Before we start harping on about DHs and smart executives, let's look at some telling numbers.
Year Average AL Team Salary Average NL Team Salary
2009 $93.3M $84.3M
2008 $96.9M $83.1M
2007 $92.8M $73.7M
2006 $83.7M $72.4M
2005 $75.5M $70.9M
(I got each year from a different source, but in any given year, the AL & NL numbers are from the same source. They're not exact, there are different ways of measuring this, I know, I know. Take them as rough estimates, I don't care.)
The obvious reaction to this is: well of COURSE the AL does better, their average salaries are noticeably higher. And certainly yes, that does give them an edge. I'm going to do some crude analysis to show you that it does partially explain the results.
Here's another table. This one shows the difference in average salary computed from the info above, and, courtesy of FanGraphs, the price of a marginal win, from 2005-2008.
Year Salary Difference Price of a Marginal Win
2008 $13.8M $4.5M
2007 $19.1M $4.1M
2006 $11.3M $3.7M
2005 $4.6M $3.4M
If we divide salary difference by marginal win price, we find that the difference in expected win totals between the average AL & NL teams should be 3.07, 4.66, 3.05, and 1.35, equivalent to season winning percentage differences of .019, .029, .019, and .008.
These are the differences we'd expect against average competition. In interleague play, you have above average teams playing against below average teams (in aggregate). So, the difference in winning percentages should actually be magnified a bit. This is quite a large expected difference in performance, but it still explains less than half of the AL's success. If you compute a z-score for the 2005-2008 results using this information, regardless of what assumptions you make, you're going to get something more than a couple standard deviations from the expected results. Thus, there's something else going on here.
|
| | Posted 6/18/2009 2:47 PM - 15 Views - 0 eProps - 1 Comment
- recommend
  - recs0
- share
- email
 - sent0
|
