MGL over at THE BOOK makes the point that the better team didn't necessarily win the World Series:

Who had the better team?

(snip)

I simply mean if they played against one another with roughly the same players (we’ll even say 4 starting pitchers only) for a few thousand games, who would likely win the most games?  (Actually, those are not exactly the same question, but we’ll assume it is.  Team A could theoretically be better than team B, against neutral or unknown opponents, but team A could match up poorly against team B and be a dog... 

He also points out that the results of your answer don't particularly matter, he's just attempting to make a point. The only rational response I can come up with is, "does that point matter?" Where's the line between rational analysis and an acknowledgment that the World Series is not, in fact, a few thousand games?

The conclusion from such an idea points to the removal of small sample sized playoff exhibitions, but I can't believe MGL or any of us, sabermetrician or otherwise, would actually prefer that.