instead of the same day - there is a most satisfactory feeling of getting something for nothing about the whole affair.
8. The chief event in November was that Dolphin and Plaice both went four wheel; all the main keys were now 4 wheel and none of them throw-on - the changeover was complete. This raised a new problem - what should we do when the bigram tables changed? Our previous method, on 3 wheel keys, had been to have EINS encoded at all machine positions on the W.O. and stecker for the day. By multiplying the number of positions, and therefore the work involved, by 26 the change from the 3 wheel to 4 wheels machine rendered this method impracticable. It was therefore necessary to develop some new attack.
9. The bigram table problem is essentially that of getting out messages when everything but the starting position is known. A number of devices had been made using one of four general methods (1) the EINS catalogue, dealing with all the traffic on a key simultaneously and getting out all messages (probably 60%) containing an EINS. (2) Rodding, by hand or mechanically (the Click machine, invented by Turing). This is a method working on the right hand wheel only and having the advantage of taking little longer on the 4 wheel machine than on the 3 wheel. As it needs an 8 letter crib, however, it is only applicable to a small number of messages. (3) Statistical, using the fact that the message contains a much greater than random number of E's, N's and other common letters. The Op. 20 G photographic machine, Hypo and the Arlington dudbuster both used this criterion. This method is effective and gets out a higher percentage than any other, but is slow if a large number of messages have to be broken. (4) A modified bombe, testing some word like EINS, or an urgency sign such as KRKR, on individual messages. The Op. 20 G "Grenade" and our "Jones Dudbuster" worked on these lines. Here again the method was too slow - even though the dudbuster tested 9 cribs at a time (only 1/9th of the bombe was used on an individual test).
-79-
Back to Cryptographic History of Work on the German Naval Enigma contents