Well, we're talking about convoys of merchant ships here. Three things have to be known exactly for a torpedo shot to intersect the target: target course (=angle on the bow), target speed, target range. To obtain these the sub has to make an "approach." (The target can be at any range at any angle on the horizon, when first contacted.) Here's an excerpt from Ch9 of "OEG Report 51 - ASW in WWII" at
Ibilio, page 95. (Underlined passage by me.)
View attachment 872197
FIGURE 3. Ships sunk per million miles of track in a U-boat density of 1 per million square miles, as a function of ship speed.
Also, the Japanese Type 93 torpedo, while fast, was only used on cruisers and destroyers. It was 24-inch diameter and way to long and heavy for submarines.
The entire document (linked above) is worth the read. While a faster torpedo should, in theory, be somewhat more accurate than a slower one, the other errors (listed above) override this by an order of magnitude. The U.S. sucessfully used an electric torpedeo in the Pacific in 1944 (the best year for tonnage sunk) and it had a speed of only 29 knots! Because it left no wake the subariners liked it a lot (once the bugs were worked out of it.)
Edit: Added this summary at end of section in document:
Edit 2: The Germans were worried about the speed of the Victory ships, as recorded in this video. Youtube link is
Drachinifel - Walter and Electro-Boots. Go to 42:50 in the video where he's talking about a proposed Type 26W U-boat (Walter-Propulsion) that would be able to use the standard "follow and end-around" approach method because of its greater speed.