I think it explains a lot of things nicely, as a good theoretical framework should. But I'm also a little cautiously skeptical about it, because you're working backwards, without benefit of any real-time experimental data, so it's almost impossible to keep "present-day bias" out of it.