170 | 16 THE TESTERy
and from these decrypts we knew the enemy’s strategic decisions and their overall thinking.
Kesselring was pushed steadily northwards by the Allies until the German Army was out of
the way.
The fourth example came after D-Day, during a period of stalemate in France when the
battle lines got bogged down. The Allies had tried but failed to take Caen in Normandy and
the Germans were holding out resourcefully. Hitler tasked von Kluge, who was in charge of the
German forces in France, with pushing the Allies back into the sea—or at the very least holding
the line. He managed to do so for 3 or 4 weeks, and then Patton broke through in the west of
France and Montgomery in the north. Between them they managed to surround part of the
German army and took 90,000 prisoners. Von Kluge was ordered back to Berlin, but committed
suicide rather than face the furious Hitler.
It is often said that Tunny decrypts helped to shorten the European war by at least two years.
This was a war in which approximately 7 million people on average died each year—so our
decrypts saved millions of lives. Much of this success was down to the work of Bill Tutte and
the Testery. Before Tutte’s work on Tunny he had been interviewed for a job by Turing, but was
rejected; luckily he was interviewed again and taken on by Colonel Tiltman. As it happened, I
then worked in the same office as Tutte and still have a clear mental image of him staring into
the middle distance for long periods, twiddling his pencil, and covering reams of paper with
calculations. I used to wonder if he was actually getting anything done, but my goodness, he was!
Breaking Tunny
Once Tutte had opened the door to Tunny, and some broken messages had revealed its impor-
tance, we realized that we must break it on a daily basis. At that point the Testery, which had
until then been working on a somewhat old-fashioned crypto-system called Double Playfair,
was switched entirely to Tunny. At the time—the middle of 1942—the Testery consisted of a
linguist and three senior cryptographers, of whom I was one. Six additional cryptographers
were brought in as the Tunny work increased, and by the end of the war the Testery had a total
staff of 118. I estimate that we produced some 64,000 decrypted messages, often running into
many pages. In terms of sheer numbers, this is mere peanuts compared with Enigma, where
more than 100 messages were being broken every hour—day and night. But Tunny messages
were nuggets of gold.
One of the most remarkable things about the Testery was how young we all were. Peter
Hilton, for example, joined during his first year at Oxford University, when he was just 19 years
old (Fig. 3.1). Donald Michie was even younger, literally a schoolboy. I was pushing it a bit,
age-wise, because I was 20 when I joined (Fig. 16.1). Tutte was 24 when he broke the Tunny
system (Fig. 14.5). As for Turing, at 29 he was positively in his dotage. But don’t get the wrong
impression: breaking Tunny was never child’s play. It was always a challenge to get each new
day’s wheel patterns for each of the five key links we worked on, since to do this we had to break
from thirty to fifty consecutive places of a message’s ciphertext. To break into a message in two
or three different places is one thing, but to break so many consecutive places is very much
more difficult. For the first year we did everything by hand, without any machines to help us,
and we broke one-and-a-half million letters of ciphertext. That takes a lot of doing.
I will explain one of the most important methods that we used in the Testery to break Tunny
messages. In the earlier days, before the tightening up of German security during 1943, there