Great question! I missed that earlier. In all playoff games since 1990 (just to expand the sample a bit more) where a +3.5 SRS or better offense faced a +3.5 SRS defense or better, the defensive team won 52.6% of the time against a pregame Elo-expected win probability of just 49.7%. (To put it a different way, the offensive teams won only 47.4% of the time and were expected to win 50.3%.)

Since 2010, the trend has deepened: Defensive teams beat offensive teams 54.4% of the time, versus an expected W% of 49.1%. So in this quick-and-dirty study at least, there seems to be some effect by which defensive-minded teams are able to beat offensive ones in the playoffs more than expected. (Watch out, Pats!)