Moderator: mosher

Chaocipher Exhibit 6: A real-life cryptanalytic challenge

mosher
Super member
mosher
Super member
Joined: May 26th, 2009, 10:24 am

January 11th, 2015, 9:09 am #1

Back in May 2013, Crypto Forum member, Chaocipher enthusiast and author Jeff Calof traveled cross-country from California to the National Cryptologic Museum in Ft. Meade, Maryland, with the express purpose of examining the entire archived Chaocipher papers, contributed by Byrne's daughter-in-law, Patricia Byrne.

One of the jewels Jeff uncovered was an unpublished paper by Lou Kruh and Prof. Cipher Deavours. The paper presented fifty (50) short messages, all in-depth (i.e., beginning with the same machine settings). Kruh and Deavours did not disclose the key or the precise Chaocipher system used (i.e., was it enciphered with the method used for Exhibits 1&4, or the method used for Exhibit 5?). They wanted to simulate what Friedman should have received from Byrne back in 1942 if the latter hadn't been so obsessed with keeping his system's inner workings secret. This exhibit has been named "Exhibit 6" for reference purposes (see page 332 in John F. Byrne's Chaocipher Revealed: An Historical and Technical Appraisal, accessible for free, for Friedman's official request for material from Byrne).

Here's a taste of what the challenge messages look like:

TIHUL RZNXN SDGQL MYGNU QQAXF H
OYRBQ NNZEG ZECMZ MOMKO AMEBL HB
XANQD WXZST SJMLR XNSLC YPDJD UO
XANQD WXOAO JZSYM OXEQE LAC
XMIRE DOHSJ HHSNQ QFHZL LHZMH G
TCYXM XFROW ACWTY QEVMF ITXTO A
NHCDQ DEGRG OMQBD RKBJX HRKQN LPOOD
TIYJD ICMSR PTVHB EUSQD KVYIT RWDL
TCQPK EGAUN TKMWK XNBNO GDUTR
BCQXD OGMPE MKBLT YUAUQ DPTJR JZH
. . .
JIYSS COITP XYJNP QZZOZ MIZOM
DVIBX GVFQP QIQXT OMDAR MPBYG KXTW
IDAYS NYRVC BBBKU UUDGZ WABAA
TCRVF CCAER NDIXR OTHJX XZTGG QDSTB FENB
YHAPC NUQLS MHRBY MVDKR NDNMC PDQN
MASXQ WVXWA UAOIV BMTIX YWONN
DVINS PRNDL LBUKC RFEFS METGM JNLNT XHGZ
TCQNO QUTHK TVGHA LVAJS CIUSQ MQ
TCBXI SZAOK OYQNT PKQOD JKUKP ZFK
BKRBO COBBL KYOQP IINKF EUYEN JRGMD


Jeff Calof and Moshe Rubin co-authored a paper describing everything about Exhibit 6, which you can check out at:

Chaocipher Exhibit 6: Real-life cryptanalytic challenge

This windfall affords all cryptanalysts (including governmental ones! :) ) a unique opportunity to test their mettle and tackle an honest-to-goodness historical cryptanalytic challenge. Given the plethora of techniques known today (e.g., the use of computers, generic algorithms, hill-climbing, differential cryptanalysis) this should revv up the juice of cryptanalysts the world over. If you've ever wanted to try your skill at solving a real-life cryptanalytic challenge, here's the opportunity to flex your muscles.

Solving this in-depth message challenge, or not solving it, would present an important statement regarding the security of the Chaocipher crypto-system.

Good luck!

Last edited by mosher on January 14th, 2015, 12:13 pm, edited 2 times in total.
Quote
Like
Share

mosher
Super member
mosher
Super member
Joined: May 26th, 2009, 10:24 am

January 13th, 2015, 8:11 am #2

If we assume Kruh and Deavours used the same enciphering method as Exhibit 5 (i.e., the plaintext disk is determined by a finite and periodic order of right/left disk), the challenge in solving this Exhibit is to determine four components:
  • The initial order of the left alphabet
  • The initial order of the right alphabet
  • The juxtaposition of the left and right alphabets
  • The order with which the plaintext alphabet is determined
Sorting the fifty messages alphabetically shows numerous initial repetitions:

AHPEV CEUYH RTANP GKZIR TS
BCQXD OGMPE MKBLT YUAUQ DPTJR JZH
BKRBO COBBL KYOQP IINKF EUYEN JRGMD
BPAEX PQVQV WAVDG EBESZ XJHAK FZ
DTGFW TSCNI DOAVR LCPAE AIXRQ YM
DVHPA XKJZN VSDCQ CJLEG AWONP LNIXG
DVIBX GVFQP QIQXT OMDAR MPBYG KXTW
DVIIX IXIPH BPZXQ XTMBG NEQ
DVIIX LKBKH LUJLY NIVOY AGULA ZZY
DVIIX LLUBE LWWQY LMFOD QQCSK KHMQ
DVINE VMEMJ JMCFR OJMQH HBVGV LCBGD LBD
DVINS PRNDL LBUKC RFEFS METGM JNLNT XHGZ
DVIQE GMDMT RTIEM BACBT WQYUU GTXXN YX
DVIUT QJXIO ZVMRW BLCYH QC
DVRHS LOCDS VJQXC FPAZT OFV
IDAYS NYRVC BBBKU UUDGZ WABAA
JISRX QMPQI QGNSJ GDZFQ MEDKC
JISUA CDUPG EZPKS VXUOE VWBBS BVSA
JIYNE WKTIS MLDKV SFZGH VZRHV DBBXJ Y
JIYSS COITP XYJNP QZZOZ MIZOM
KMPNS TOABI TQGSU JMXRZ KKNJN UXV
MASXQ WVXWA UAOIV BMTIX YWONN
MSQEJ KMEYK JQSXG QJDHB YYEKQ NHX
NHCDQ DEGRG OMQBD RKBJX HRKQN LPOOD
OKLAE CKPKN LTSGY AISUL NIFFR Y
OKLBX QNFLP YDIKL OHMTH PWUWM SH
OKLNC FSRRC ZRIRC BWXBW CO
OYRBQ NNZEG ZECMZ MOMKO AMEBL HB
SYGYE ZEEJJ VGBYF BJGNC BMNRU XQPUN
SYMEZ WGNPH QEPLZ LSZIT DPYIT
TCBXI SZAOK OYQNT PKQOD JKUKP ZFK
TCCTS OLEBN BFHID BURNI IVPPZ UC
TCGHV HLNNJ IDIKU FNUNU IHXNR APTC
TCQNO QUTHK TVGHA LVAJS CIUSQ MQ
TCQPK EGAUN TKMWK XNBNO GDUTR
TCRBD XCAPM NTPHK HKNFZ JNHIK L
TCRVF CCAER NDIXR OTHJX XZTGG QDSTB FENB
TCYCS WGOOH UJRDB EGFXP TTRHU CJLY
TCYXM XFROW ACWTY QEVMF ITXTO A
TIHUL RZNXN SDGQL MYGNU QQAXF H
TIYJD ICMSR PTVHB EUSQD KVYIT RWDL
TLQYI LXKSU WMWAE JUZHI WPQNT TI
TPUSB SWAJZ SPMBH HTSFO RXYED DA
TPZEY GGMVA XZKSQ LYBEB TXTRM YEDHK
VISSL XWQIE DRIPN MZONC OWNMT Y
XANQD WXOAO JZSYM OXEQE LAC
XANQD WXZST SJMLR XNSLC YPDJD UO
XMIRE DOHSJ HHSNQ QFHZL LHZMH G
YHAPC NUQLS MHRBY MVDKR NDNMC PDQN
YIMEH ZZRVC ZAHYZ VYFZP JXAFS GQFE

The beginnings of "DVI", DVIIX", DVIN", and "DV" look like good candidates for common words beginning with "TH": "THE", "THERE", "THEN", etc. It might not solve the messages immediately, but William F. Friedman made use of message beginnings to break into the Wheatstone Cryptograph (for a solution of the Wheatstone using similar beginnings, see WFF's "Several Machine Ciphers and Methods for their Solution", Riverbank Publication No. 20, Geneva, IL, Riverbank Laboratories, 1918, pp. 6-36). If one assumes, for example, that "DVI" (ct) is "THE" (pt), this theoretically reduces the possible candidates for the four components mentioned above. Can a cryptanalyst leverage this knowledge?
Quote
Like
Share