Here's a printable card summarizing the Dick O'Kane method. Torpedo speed, while part of vector analysis, is not part of the Dick O'Kane method. The theory of the Dick O'Kane method is to eliminate human error. A large source of human error is to make a solution based on the wrong torpedo speed. The TDC knows what torpedo you have loaded and what speed it will go based on the switches. Why would you want to manually track all that stuff. Just let the stupid machine do it!
As you can see I'm not even worried about whether the torpedo track angle is exactly 90º. Any shooting at gyro angles between +-20º is considered straight shooting anyway and you'll be well under +-5º if you just choose 10º for target speed under 15 knots and 20º for speeds over 15 knots.
I also never shoot from as far away as 1500 yards unless it's my first shot at an escorted convoy, when I'll shoot from 3,000. I'm looking for a range between 500 yards and 750 yards when I'm shooting one or two unescorted merchies. And NEVER USE SLOW SPEED! Your accuracy is directly proportional to your torpedo speed. Half as fast means half as accurate.
So Dick O'Kane uses the TDC to calculate the gyro angle. You choose a lead angle arbitrarily, based on a rule of thumb because it just isn't worth the bother to be precise. The TDC compensates anyway. People do stupid easy stuff. Machines do fussy precise stuff. The combination gives you hits.
In the video he didn't show everything he did. For instance you can't verify that he got the speed he claims by watching his actions. It's like he set up the attack and phoned in the procedure. Can't really learn how to shoot that way.
Missing astern is usually torpedo doors not opened or bad target speed determinaton.
Missing ahead is using fast torpedoes and doing a calculation for slow ones (using the TDC makes that error impossible) or bad target speed determination.
If you enter starboard AoB instead of a correct port AoB or versa visa you'll miss astern.