Here the
SAFEST reply
is rarely the
f > n
SANEST