Here is the fire control manual from the 1950s for fleet boats:
http://www.hnsa.org/doc/attack/index.htm
Thge optimum track angles are shown in plate XVII and plate XVIII. What calculations that graph is based on isn't shown. It appears from the graph though, that for a zero gyro angle shot, the maximum deflection angle you can get is the best.
That graph is a theorhetical example as it's for a zero gyro angle shot with a torpedo run length of 1,000 yards, so it's not something that you could manouver for. So, trying to read that graph as it is is quite hard.
If though you are shooting from behind, the speed of the torpedo will be reduced in relation to the target, which will mean that the target and torpedo have more chance to converge, as the torpedo will spend longer travelling across the target's course so the effective target length is larger. The opposite of the Cromwell method, where the target speed adds to the closing speed.
BUT! That graph is misleding, as you cannot directly compare a shot with a track of 45 degrees and a shot with 120 degrees, as you will need to be in completely different positions to get the same theorhetical 1000 yard run and zero gyro angle. Obviously you will need to be much closer if shooting from behind than if you are shooting from ahead to achieve a 1000 yard run length for both shots and the range difference in the two firing positions will increase the faster the target is.