The problem of minimizing mean-square error when digital data is encoded into the elements of a group code via an encoding map a and the received elements of the group code are decoded via a decoding map

is considered. This problem has been solved under the restriction that

. It is shown that the same solution still obtains when this restriction is removed.