At the time, it was ascribed to a conversion error between metric and inches. I think I read later it was something else, which eludes me.
Zimmerman discusses that whole thing.
In brief;
The device they used to test the mirror during grinding, was mechanically mounted with a small gap that wasn’t taken into account in the calculations. Fortunately,(or not) it was a consistent-error during grinding, so the whole mirror was ground to the exact wrong prescription.
It was a failure in the metrics method chosen......old school optical engineer didn’t want to use new tools