Hut 8, and the war beneath the waves 

The German navy took cipher security more seriously than the army or air force. Naval Enigma drew its rotors from a larger set of eight, its indicator procedure was more careful, and its operators were less prone to the lazy habits that gave Hut 6 its openings. For most of 1940 the naval traffic was effectively closed.

Turing ran the section attacking it, Hut 8, with Hugh Alexander, the chess champion, later taking over the day-to-day management. Their statistical method, known as Banburismus after the town where the long sheets of paper were printed, used the overlap between messages to narrow down which rotors were in use before the Bombes were committed, saving machine time that was desperately scarce.

But mathematics alone was not enough. Hut 8 needed key material, and key material had to be physically taken. In 1941 British forces seized documents from German trawlers off Norway and from the submarine U-110, whose capture yielded settings and cipher papers. Those pinches, as they were called, gave the codebreakers windows of readable traffic, and readable traffic gave them the cribs they needed to keep going after the window closed.

Then, in February 1942, the U-boat fleet moved to a four-rotor machine on its own network, called Shark at Bletchley. The extra rotor multiplied the search space beyond the reach of the existing Bombes, and the section went nearly blind for most of the year. Shipping losses in the Atlantic rose sharply. The blackout ended largely because of another pinch: in October 1942 three sailors from HMS Petard boarded the sinking U-559 in the Mediterranean and passed out its short weather cipher documents. Two of them, Tony Fasson and Colin Grazier, went down with the submarine. Their material gave Hut 8 the entry back into the four-rotor traffic.

What the intelligence bought was not a single dramatic victory. It was routing. If you know roughly where the wolf packs are gathering, you can send a convoy somewhere else, and a convoy that is never attacked leaves no story behind. Historians generally credit signals intelligence, together with radar, better escorts and long-range aircraft, with shortening the Atlantic campaign considerably rather than winning it single-handed.

Every reroute carried a risk. Divert too obviously, sink too many submarines too precisely, and the Germans would ask why. Protecting the secret was as much of an operational problem as producing it.

That problem did not end in 1945. It lasted another thirty years.