Next: Work Description
Up: Stage 8
Previous: Stage 8
Finally a stage where it is obvious what crypto was used to
generate the ciphertext.
In principle, we could have attacked this stage immediately, but being
placed as the 8th problem we thought that it would be too difficult.
After some skirmishes with Stage 6 without any success, some team members
turned their attention to this problem, and it was pretty much solved
in parallel with Stages 6 and 7.
So what do we know, and what information do we need?
It is clear from the illustration in Stage 8 that the ciphertext was
encrypted with a three-wheel Enigma for which the contents of the three
wheels as well as the reflector are known.
This means that there are four classes of unknown parameters:
- The direction the wheels rotate
-
It is not clear from the figure in what direction the wheels rotate.
- The initial wheel positions
-
For each of the three wheels, there are 26 initial rotations possible.
- The ring settings
-
The first two wheels contain two rings.
Remember that the wheels are rotated after each encrypted character in
an odometer-like fashion:
The first wheel is always rotated, the second wheel every 26 rotations of
the first wheel, and the third wheel every 26 rotations of the second
wheel.
The rings determine the point of the first rotation of a wheel:
If the ring on the first wheel is reached after 10 rotations, then the
second wheel is rotated after 10, 36, 62, ...rotations.
For the first two wheels, there are 26 possible ring settings, but we
can actually disregard some placements on the second wheel as the ciphertext
is shorter than 676 characters long.
- The plugboard connections
-
The plugboard contains a set of pairs of letters which are exchanged for
each other: If E and Z are connected, then every in- or out-going signal
which is an E is mapped to a Z, and vice versa.
According to Singh, there were at most six pairs of exchanged letters.
This is what we deduced from the problem formulation.
In fact, we also introduced another degree of freedom as one
team member interpreted the wheel transpositions differently.
This later doubled the CPU time needed to find the solution.
Next: Work Description
Up: Stage 8
Previous: Stage 8
solvers@codebook.org