establish the true distance (as he would be able to do in this case) and at worst to be able to indicate some portions as more likely than others. This latter possibility is worth going into a little more fully. Suppose we tested Q = B + 9; then we should get
Now .Y.M. is an impossible beginning for a message and therefore, if we are really confident that the clear text for TYB is correct, Q = B + 9 cannot be right. Again, suppose that for Q = B + 5 we had not got conclusive evidence as in my example (which is far more favourable than the average case) but had got, say, ...... N....E...ER....S.... . Although it may be impossible to fill in the text of this nevertheless these are all very common letters and consequently increase considerably the chance of the position being right. It is quite easy to score consequences of this kind mathematically in the same units as were used in the rest of Banburismus and when we tried a crib against another message in this way (the process was known as a "slide" to us) the results were scored and entered on the deciban sheets.
9. The example may give a false impression that the whole process was quite easy. It was possible to work a shift without a single success and there was a great difference in the results obtained by the good and not cribsters; judgment as to what was and was not worth tackling, speed of work, good cooperation with the Banburists and ability in building up depth from a slender start were all very important and gave a great deal to scope to the individual. A process known for no very good reason as "Babbageismus" which consisted in taking a message about which a great deal was known and sliding it against all messages whose trigram had the first letter the same - which might involve a very rapid examination of hundreds of positions - was also used with effect by the better cribsters.