from Amazon
Thursday, April 29, 2010
Superiority of AL over NL
Matt breaks down the talent by service class and concludes that:
Usually the numbers are not this crisp, but they all seem to point to the American League’s superiority coming from spending more money on young talent, while gaining no advantage from spending more money on free agents. This appears to be even stronger evidence of the repeated prisoner’s dilemma nature of the draft. There may not be any need for Major League Baseball to correct for this discrepancy, but it is certainly worth determining why it exists. These numbers appear to do exactly that.
A good breakdown.
***
Though this was not part of his main point, Clay is wrong here:
In this article, I am less focused on how much better the American League is than the National League. Clay Davenport can do that far more accurately. His league adjustments, which are the primary difference between WARP1 and the newer WARP2 and WARP3 indicate that National League teams had an average WARP3 of 40.2 over the last two years, while American League teams had an average of 42.9. Thus, he has estimated that the talent level appears to be about 2.7 wins higher per team. That seems reasonable enough.
This is very unreasonable. The gap between an AL and NL team is only 2.7 wins? I use a conservative +.05 wins per game gap, which is 8 wins. What’s the real answer? Well, MGL did a fantastic three-parter for THT a few years back. I’ll do it the very simple way.
In 3208 interleague games, the average AL team wins .522 games against the average NL team (which obviously wins .478 games). That puts the gap at .022 wins per game, or 3.6 wins. Sounds like what Matt is reporting Clay’s data is showing.
Here’s the year-by-year gap:
(1) 1997
1 1998
(7) 1999
3 2000
0 2001
2 2002
(2) 2003
3 2004
11 2005
13 2006
11 2007
15 2008
9 2009
You can certainly make the case that there’s been a huge shift sometime around 2005. This is still sample data, and I’m cherry-picking, so you need regression.
From 2005-2010, based on 1260 games, the average AL team has won .574 of inter-league games (and .426 for the NL). This means that if you stick the average NL team of 2005-2010 into the AL, they would win .426 games in the AL (as opposed to .500 in the NL). That is a huge 12 win gap.
That’s why I’m using 8 wins, as halfway between the historical data (1996-2010) of 4 wins and the recent data (2005-2010) of 12 wins. If I did a weighted average, and used each season as 20% more than the previous, I get 8.5 wins. If I use 10% instead, I get 6.5 wins. Whatever it is, it’s far more than 2.7 wins.
***
Reference tool: B-R.com
Recent comments
Older comments
Page 1 of 344 pages 1 2 3 > Last »Complete Archive – By Category
Complete Archive – By Date