Changes

There are other things that one must do.  One must avoid using stereotyped language.  If every message starts with "Today's orders are to go to location XYZ and patrol for surface shipping", it becomes easier to decrypt.  The people at Bletchley were quite good at guessing the German's writing style, and could make guesses (called "cribs") of what they might be talking about.  A common way to avoid intentional or unintentional use of stereotyped language is to start each message with a nonsense sentence.  This is similar to what at present is called "word salad" by email spammers to evade spam filters.

There are other things that one must do.  One must avoid using stereotyped language.  If every message starts with "Today's orders are to go to location XYZ and patrol for surface shipping", it becomes easier to decrypt.  The people at Bletchley were quite good at guessing the German's writing style, and could make guesses (called "cribs") of what they might be talking about.  A common way to avoid intentional or unintentional use of stereotyped language is to start each message with a nonsense sentence.  This is similar to what at present is called "word salad" by email spammers to evade spam filters.

The Nazis used word salad, but they also made a costly mistake.  They routinely ended each message with "HH", for "Heil Hitler".  This foolish bit of vanity was helpful to the British.

The Nazis used word salad, but they also made a costly mistake.  They routinely ended each message with "HH", for "[[Heil Hitler]]".  This foolish bit of vanity was helpful to the British.

There is another problem.  The movement of the rotors meant that simple letter frequency analysis, which plays an important part in solving cryptograms in magazines ("This message has a lot of m's; let's guess that m corresponds to e.") won't work.  But one can still do frequency analysis on the letters in a corresponding position in the messages, since they all used the same key.  That is, if today's messages often have m appearing in the 15th position, one could guess that those were e's in the 15th position, since everyone's rotors would have been in the same position at that point.  While utilizing this information would have been much harder than ordinary substitution ciphers, it was still a weakness.

There is another problem.  The movement of the rotors meant that simple letter frequency analysis, which plays an important part in solving cryptograms in magazines ("This message has a lot of m's; let's guess that m corresponds to e.") won't work.  But one can still do frequency analysis on the letters in a corresponding position in the messages, since they all used the same key.  That is, if today's messages often have m appearing in the 15th position, one could guess that those were e's in the 15th position, since everyone's rotors would have been in the same position at that point.  While utilizing this information would have been much harder than ordinary substitution ciphers, it was still a weakness.