Hollinger is willing to stick by his statistics. I respect him for that. If his pick is based on "point differential" then Boston will appear to him a very weak team. We know that Boston plays down to competition and we drop too many games to sub .500 teams. However Boston has a very good record against the top teams. This leads them to "look worse" than they actually are.
Miami's record against the elite teams was not too impressive, but they often destroyed the sub .500 teams by a large margin. I believe this leads them to "look better" than they actually are.
You can't measure heart, and the Celtics play with a ton of heart in the post-season. I wonder if Hollinger made any adjustments to his system based on his previous predictions and how they turned out.