The single point of failure lies in extrapolating from a test bench limited to a supposed large-scale optimization. A test on only 250 assets does not allow validation of performance for portfolios with thousands of securities, where the blast radius of an error would be enormous. This is an unacceptable risk for capital preservation, requiring much more rigorous validation before any implementation in an institutional context such as a pension fund. For example, a Swiss manager would not consider this as proof of robustness for managing allocations including complex derivatives or exotic markets.