Your models all have a first section to generate random numbers and a second to encipher with disk(s). None of the models is good suggesting
a. the random generator is wrong and the enciphering is OK;
b. the random generator is OK but the enciphering is wrong
c. both sections are wrong.
The Markov generator gets close to the incidence wave. Perhaps the enciphering system is wrong.
Have any of you tried this enciphering system:
encipher plain letter with first disk and add key to get intermediate.
encipher intermediate on second disk.
Friedmans letter to Byrne says the cigarbox model was difficult to operate. There must have been more in the box than just a couple disks. Perhaps a mechanical drive of some kind that was flimsy/unreliable. Did the first disk drive the second through some device that added the key?
Well, about that...james wrote:There must have been more in the box than just a couple disks.