The new one failed in ways it was not designed to fail. In C-compiler terms it was "undefined behavior." In Donald Rumsfeld terms it was an "unknown unknown."

The mere fact that the outcome was successful does not inherently indicate that the decision-making was safe: the O-rings "worked" for 24 missions and the foam/tiles "worked" for 111. Nevertheless there were ample warnings and close calls.

Reentry from the Moon is not routine. Re-entry speed was about 40% faster than from low earth orbit, and kinetic energy goes up by the square, so about double.