Feynman Ciphers

According to a 1987 post by Chris Cole (who was then working at Peregrine Systems) to sci.crypt on Usenet, the following three short ciphers were passed to Richard Feynman by an unnamed fellow scientist at Los Alamos (hence “The Feynman Ciphers”). The first was cracked by Jack C. Morrison of JPL, but the other two remain unbroken.

Feynman Challenge Cipher #1 (Solved!)

MEOTAIHSIBRTEWDGLGKNLANEAINOEEPEYST
NPEUOOEHRONLTIROSDHEOTNPHGAAETOHSZO
TTENTKEPADLYPHEODOWCFORRRNLCUEEEEOP
GMRLHNNDFTOENEALKEHHEATTHNMESCNSHIR
AETDAHLHEMTETRFSWEDOEOENEGFHETAEDGH
RLNNGOAAEOCMTURRSLTDIDOREHNHEHNAYVT
IERHEENECTRNVIOUOEHOTRNWSAYIFSNSHOE
MRTRREUAUUHOHOOHCDCHTEEISEVRLSKLIHI
IAPCHRHSIHPSNWTOIISISHHNWEMTIEYAFEL
NRENLEERYIPHBEROTEVPHNTYATIERTIHEEA
WTWVHTASETHHSDNGEIEAYNHHHNNHTW
(380 characters, 5 x 76 transposition cipher, start from the last position and step back 5 each time, then repeat starting from the letter one before the last position etc) --> Chaucer, Canterbury Tales:-
WHANTHATAPRILLEWITHHISSHOURESSOOTET
HEDROGHTEOFMARCHHATHPERCEDTOTHEROOT
EANDBATHEDEVERYVEYNEINSWICHLICOUROF
WHICHVERTUENGENDREDISTHEFLOURWHANZE
PHIRUSEEKWITHHISSWEETEBREFTHINSPIRE
DHATHINEVERYHOLTANDHEETHTHETENDRECR
OPPESANDTHEYONGESONNEHATHINTHERAMHI
SHALVECOURSYRONNEANDSMALEFOWELESMAK
ENMELODYETHATSLEPENALTHENYGHTWITHOP
ENYESOPRIKETHHEMNATUREINHIRCORAGEST
HANNELONGENFOLKTOGOONONPILGRIM

Feynman Challenge Cipher #2

XUKEXWSLZJUAXUNKIGWFSOZRAWURORKXAOS
LHROBXBTKCMUWDVPTFBLMKEFVWMUXTVTWUI
DDJVZKBRMCWOIWYDXMLUFPVSHAGSVWUFWOR
CWUIDUJCNVTTBERTUNOJUZHVTWKORSVRZSV
VFSQXOCMUWPYTRLGBMCYPOJCLRIYTVFCCMU
WUFPOXCNMCIWMSKPXEDLYIQKDJWIWCJUMVR
CJUMVRKXWURKPSEEIWZVXULEIOETOOFWKBI
UXPXUGOWLFPWUSCH

Feynman Challenge Cipher #3

WURVFXGJYTHEIZXSQXOBGSVRUDOOJXATBKT
ARVIXPYTMYABMVUFXPXKUJVPLSDVTGNGOSI
GLWURPKFCVGELLRNNGLPYTFVTPXAJOSCWRO
DORWNWSICLFKEMOTGJYCRRAOJVNTODVMNSQ
IVICRBICRUDCSKXYPDMDROJUZICRVFWXIFP
XIVVIEPYTDOIAVRBOOXWRAKPSZXTZKVROSW
CRCFVEESOLWKTOBXAUXVB

Note: I found these ciphertexts on John’s Javascript Secret-Code Systems page.

7 thoughts on “Feynman Ciphers

  1. Anne-Lise Pasch on January 25, 2013 at 12:09 pm said:

    The first Feynman is a modulo 384 transposition cipher? I’m pretty certain it’s a 5×76 Horizontal / Vertical route transposition cipher. Can it be both?

  2. Anne-Lise: as far as I know, they’re both semi-mathematical ways of describing the same thing. Which is “start at the end letter, and step backward 5 characters at a time (one pass at a time) to reveal the hidden message“. 🙂

  3. Anne-Lise Pasch on January 25, 2013 at 10:38 pm said:

    *g* Right then. I need to learn more math! :>

  4. http://www.austininc.com/SciRealm/CodeSystems.html

    Hi, just letting you know that the link to my site changed – AOL dropped hosting awhile back

  5. Anne-Lise: revisiting this page, I think your description is a bit more accurate (the maths is sort of correct, but not particularly helpful), and that Feynman Cipher #1 would indeed be better described as a 5×76 transposition cipher. Text updated accordingly, thanks! 🙂

  6. Pingback: A quick look at the Feynman Ciphers... -Cipher Mysteries

  7. Graeme Cole on March 11, 2018 at 12:48 am said:

    I’ve been looking at the second Feynman cipher, and I’ve noticed that the letter U appears 23 times but only ever follows one of eight letters: X, J, W, M, L, D, T and I.

    This small number of distinct predecessors for a letter with that high a frequency is extremely unlikely to happen by chance, if the letters of the ciphertext were in a random order. I wrote a script to generate and check a million random permutations of the ciphertext, and the letter U had eight or fewer distinct predecessors in only 22 of these permutations.

    This suggests to me that there isn’t any fancy transposition path required for the second cipher – the letters of the ciphertext are already in the right order, or at least any transposition should preserve the relationship between the Us and their predecessors. It’s plausible that you might have to read it right to left, but not write it in a rectangle and read it vertically, for example.

    As a curious side note, I did wonder if those eight letters X J W M L D T I might represent Roman numerals, with J representing I, W representing V, and the solitary TU perhaps a miscopied JU, but maybe that’s too tenuous. Certainly it doesn’t seem to have helped me make any headway into solving it.

    What do you think?

Leave a Reply

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.