Sequence con

Here’s a sequence transposition con (primer 01613):

EAAAT SHTTI NDNTA SIFHE SDCAE WNLHA BHISF GREEN PNIBT HDEDT RIATT ETFSS IGIAD TIMEE RUTNI ARAYR NAOEA MSPRA RTEWD MAEIL HNRHA HNESS

Crib: ZOKMO

3 thoughts on “Sequence con”

  1. My hill- climber solved it without the crib. Plaintext continues in caesar-shifted form:

    HQNUXGDMJXMTZYJI

    Incidentally, to save my work, I brought up Google search and entered the address: “doc.new” and a new Google doc popped up that I could paste the solution into.

  2. I tried to devise a way to solve without the primer or a crib. I wrote a brute force program but it takes much too long to do all 99999 primers and all 3628800 keys. I’m leaving out 00000 as a primer since it would be plaintext. I wrote a converging keys solver that cuts the keyspace down a great deal by limiting keys to ten-letter words, assuming that’s how the key was generated. That’s about 10000 keys. Although this seems doable, it still takes too long. I also found that the keys didn’t converge. If you change a single digit on the primer it totally changes the encryption/decryption so getting “close” to correct is useless. It would only work if the random selection of the primer or the keyword happened to hit the correct one. There are other problems with this approach, too. There are two ways to render a word to a 10-digit key (0 low or high). The description in the article shows it as high. The resulting encryption/decryption is totally changed depending on which is used. The article also shows the key being generated using a two-word phrase. That also expands the keyspace greatly. The bottom line is that I was unsuccessful. I found that it takes about 13 seconds to brute force all primers against a single key. I think the decryption process is too slow that direction because of the overhead generating a new primer string, counting the digits, and making the “box” array for every decryption.

    1. I found the same thing — hill-climbing doesn’t work on primer digits. And since the letters are probably already in standard plaintext frequencies, the Chi-square test you can use on Gromark ciphers doesn’t help.

Leave a Reply

Your email address will not be published. Required fields are marked *