Just reread Bernstein this week and used it to force me to take a relook at my portfolio results over last 7 years and compare it to multiple benchmarks. Always a good eye opening experience.
My major critique is that he tends to be limited in how he looks at asset classes. All you need to do is to look at David Swensen's long-term experience at Yale to see that a broader view at asset allocation (e.g., Relative value / hedge funds, timber, private equity ) reduces volatility and increases return.
Interesting when looking at the Asset Allocation book written in 7-8 years ago about how right he was in terms of relative attractiveness of different assets though.