occasionally we possessed for a time what seemed to be the crib to end all cribbing. For some 3 months in the summer of 1942 Boulogne sent a weather message which began ZUSTANDOSTWAERTIGERKANAN and, if I remember rightly, it only failed twice during that period. On the whole a crib that had more than 2 or 3 basic forms was little use except for getting out paired days on known wheel orders or for depth cribbing: most cribs produced "horrors" from time to time which we classed as "other forms" and made no attempt to allow for them as a possible form of a crib, but when assessing a crib it was of course necessary to take into consideration the frequency with which "other forms" were tending to appear.
The crashing property of the machine was an essential element of all cribbing and most especially of straight cribbing. When writing a wrong crib under a portion of cypher text, each letter of crib had 1 chance in 25 of being the same as the letter of cypher text above it and of thus proving by a 'crash' that the crib was wrong. If therefore a crib had 2 good forms, each about 30 letters long, there was an odds on chance that the wrong form would crash out; in this case the remaining form is of course left with a heavily odds on instead of an approximately evens chance of coming out. If the good forms crashed out and only some rather poor form went in, it proved to be bad policy to believe the poor form to be correct although on the basis of mathematical calculation it might appear to have a reasonably good chance. It was the ability to assess this type of problem which distinguished the good cribster. It was impossible to become a good cribster until one had got beyond the stage of believing all ones own cribs were right - a very common form of optimism which died hard.
The pleasure of straight cribbing lay in the fact that no crib ever lasted for very long; it was always necessary to be looking for new cribs and to preserve an open mind as to which cribs offered the best chance of breaking a day. A crib which lasted for 2 months was a rarity; most cribs gradually deteriorated and never recovered until eventually we only recorded them every 2 or 3 days. To give up recording cribs because they were bad was a fool's policy; it unquestionably paid to keep a record of anything that might one day assist in breaking. Time and again, when a good crib died, we were thrown back onto a reserve at which we would have turned up our noses a week before, only to find that the reserve was quite good enough to enable us to break regularly.
The perennial mortality of cribs was undoubtedly to a considerable extent the result of the work of the German Security Service to whose work as crib hunters we must in all fairness pay tribute. The information recently received that they kept