I don’t think the numbers in your link match the description you give here.
1999 - YANKEES v. Braves - 16.0
2000 - YANKEES v. Mets - 12.4
2001 - Yankees v. DIAMONDBACKS - 15.7
2002 - ANGELS v. Giants - 11.9
2003 - Yankees v. MARLINS - 13.9
2004 - RED SOX v. Cardinals - 15.8
2005 - WHITE SOX v. Astros - 11.1
2006 - Tigers v. CARDINALS - 10.1
Some observations:
1. The three highest ratings earned by World Series matchups in this period featured either the Yankees or the Red Sox. The three lowest rated series featured neither of those teams.
2. The mean rating of series featuring the Yankees was 14.5. The mean rating of all other series was 12.2. The ratings of the three series containing two teams of less than national prominence were 11.9, 11.1, and 10.1.
3. Interestingly, the one time a Yankees series pulled less than a 13.0 was when they played the Mets in 2000.
What does this suggest? That your observation is dead wrong. Ratings do vary according to who is playing. When relatively low-profile teams are on both sides, ratings are dismal. However, when relatively high profile teams are on both sides, ratings are similarly poor.
This implies to me that from a ratings standpoint, the absolute best scenario for MLB is to have a high profile team (like the Yankees, Red Sox, or Mets, among others) play against a low profile team. The high-profile team brings in its national following, and brings in casual fans who have only heard of five teams in their lives. But it also brings in those who want to root for the low-profile team, because they hate the high-profile team.
White Sox v. Astros didn’t sell because the viewers there were fans of the two teams, and really hard core fans who will watch anything baseball. Yankees v. Diamondbacks sold because you had Yankee fans, Diamondbacks fans, casual fans who really only know the Yankees, AND all the people in cities across the country who just wanted to see the Yankees lose.
Yankees v. Mets still drew reasonably well because both teams have large fanbases and appeal to very casual fans, but not as well as Yankees v. Diamondbacks because the people who would tune in just to root against the Yankees for being the Yankees are by and large people who root against the Mets for being the Mets.
So I guess Boston v. Colorado would be optimal from a network standpoint. Given the national prominence of the Red Sox, the team’s large native fanbase, the people-rooting-against-the-big-bad-famous-team factor, and the good story that is the 2007 Rockies, this could match 2001 and 2004, ratings-wise.