Imagine you have knowledge of an investment signal with positive excess return, commonly referred to as alpha.

Now imagine that the entire world suddenly gained knowledge of this signal and was able to trade it on the same terms as you.

Do you believe this signal would still provide an excess return for you?

No?

Then why would someone believe that the Black-Litterman model and mean-variance optimization enable them to build portfolios better than the market?

If it is obvious and easily accessible, there is no excess return.

Similar to investment signals, good portfolio construction is about getting many small and subtle things right.

When you use Black-Litterman and mean-variance, you are getting most of them wrong.