[JoGu]

Cryptology

Breaking a Polyalphabetically Encrypted Text by Column Analysis

a7Hzq .#5r<
kÜ\as TâÆK$
ûj(Ö2 ñw%h:
Úk{4R f~`z8
¤˜Æ+Ô „&¢Dø

Columns of Ciphertext

The columns of ciphertext are arranged according to the supposed period 7:

AOWBKNL RMGEAMY CZSFJOI YYVSHYQ PYKSONE MDUKEMV EMPJBBO
AYUHCBH ZPYWMOO KQVZEAH RMVVPJO WHRJRMW KMHCMMO HFSEGOW
ZKIKCRV LAQDXMW RMHXGTH XMXNBYR TAHJUAL RAPCOBJ TCYJABB
MDUHCQN YNGKLAW YNRJBRV RZIDXTV LPUELAI MIKMKAQ TMVBCBW
VYUXKQX YZNFPGL CHOSONT MCMJPML RJIKPOR BSIAOZZ ZCYPOBJ
ZNNJPUB KCOWAHO OJUWOBC LQAWCYT KMHFPGL KMGKHAH TYGVKBS
KLRVOQV OEQWEAL TMHKOBN CMVKOBJ UPAXFAV KNKJABV KNXXIJV
OPYWMWQ MZRFBUE VYUZOOR BSIAOVV LNUKEMV YYVMSNT UHIWZWS
YPGKAAI YNQKLZZ ZMGKOYX AOKJBZV LAQZQAI RMVUKVJ OCUKCWY
EALJZCV KJGJOVV WMVCOZZ ZPYWMWQ MZUKREI WIPXBAH ZVNHJSJ
ZNSXPYH RMGKUOM YPUELAI ZAMCAEW ODQCHEW OAQZQOE THGZHAW
UNRIAAQ YKWXEJV UFUZSBL RNYDXQZ MNYAONY TAUDXAW YHUHOBO
YNQJFVH SVGZHRV OFQJISV ZJGJMEV EHGDXSV KFUKXMV LXQEONW
YNKVOMW VYUZONJ UPAXFAN YNVJPOR BSIAOXI YYAJETJ TFQKUZZ
ZMGKUOM YKIZGAW KNRJPAI OFUKFAH VMVXKDB MDUKXOM YNKVOXH
YPYWMWQ MZUEOYV ZFUJABY MGDVBGV ZJWNCWY VMHZOMO YVUWKYL
RMDJPVJ OCUKQEL KMAJBOS YXQMCAQ TYASABB YZICOBX MZUKPOO
UMHEAUE WQUDXTV ZCGJJMV PMHJABV ZSUMCAQ TYAJPRV ZINUONY
LMQKLVH SVUKCWY PAQJABV LMGKUOM YKIZGAV LZUVIJV ZOGJMOW
VAKHCUE YNMXPBQ YZVJPQH YVGJBOR BSIAOZH YZUVPAS MFUKFOW
QKIZGAS MMKZAUE WYNJABV WEYKGNV RMVUAAQ XQHXKGV ZHUVIJO
YZPJBBO OQPEOBL KMDVONV KNUJABB MDUHCQN YPQJBAH ZMIBHWV
THUGCTV ZDIKGOW AMVGKBB KKMEABH QISGODH ZYUWOBR ZJAJETJ
TFUK

Frequency Analysis of the Columns

     Column:   0    1    2    3    4    5    6

   Letter A:   4   10    8    5   17   26    0
   Letter B:   4    0    0    3   11   26    6
   Letter C:   3    7    0    6   13    1    1
   Letter D:   0    6    3    6    0    2    0
   Letter E:   3    2    0    9    7    5    6
   Letter F:   0    9    0    4    5    0    0
   Letter G:   0    1   17    3    7    4    0
   Letter H:   1    8    9    6    6    1   15
   Letter I:   0    4   14    1    4    0    8
   Letter J:   0    6    0   32    3    5    9
   Letter K:  15    6    8   28    9    0    0
   Letter L:   9    1    1    0    5    0   10
   Letter M:  14   29    4    4    7   11    4
   Letter N:   0   18    5    2    0   10    4
   Letter O:  10    3    2    0   30   17   10
   Letter P:   3   10    5    1   15    0    0
   Letter Q:   2    5   14    0    3    6   10
   Letter R:  11    0    6    0    2    4    6
   Letter S:   2    5    4    4    2    3    5
   Letter T:  12    0    0    0    0    6    3
   Letter U:   6    0   35    3    5    5    0
   Letter V:   6    5   13   10    0    6   34
   Letter W:   6    0    3   11    0    8   15
   Letter X:   2    2    2   11    8    2    3
   Letter Y:  27   13    9    0    0    7    7
   Letter Z:  22   12    0   13    2    6    5

Observation: The cipher uses strongly related alphabets.

Evidence: The »double peak« in each column catches the eye. This regularity becomes even more conspicuous, if we shift the frequency distributions to congruent positions with the double peak at the end (now written in rows):

4  4 3 0 3 0 0 1 0 0 15  9 14 0 10 3 2 11 2 12  6 6 6 2 27 22
3 10 5 0 5 0 0 5 0 2 13 12 10 0  7 6 2  9 1  8  4 6 6 1 29 18
3  2 9 0 8 0 0 3 0 0 17  9 14 0  8 1 4  5 2  5 14 6 4 0 35 13
0  4 2 0 1 0 0 4 0 3 10 11 11 0 13 5 3  6 6  9  4 3 6 1 32 28
3  2 2 0 5 0 0 8 0 2 17 11 13 0  7 5 7  6 4  3  9 5 7 0 30 15
1  2 5 0 4 1 0 5 0 0 11 10 17 0  6 4 3  6 5  6  8 2 7 6 26 26
3  7 5 0 6 1 0 6 0 0 15  8  9 0 10 4 4 10 0 10  6 5 3 0 34 15

A graphical representation shows this effect even more clearly—the first diagram showing the frequencies in the unshifted columns, the second diagram, the »fitted« frequency graphs.

[Häufigkeiten roh]
[Häufigkeiten zurechtgerückt]

Shifting the Alphabets

We adjust the columns 1 to 6 in congruence with column 0. That is the letters of column 1 are shifted by 12 positions in the alphabet: M becomes Y, N becomes Z, O becomes A and so on. In the same way column 2 is shifted by 4, column 3 by 15, column 4 by 10, column 5 by 24, column 6 by 3.

This procedure results in a text that is supposed to be monoalphabetically encrypted:

AAAQULO RYKTKKB CLWUTML YKZHRWT PKOHYLH MPYZOKY EYTYLZR
AKYWMZK ZBCLWMR KCZOOYK RYZKZHR WTVYBKZ KYLRWKR HRWTQMZ
ZWMZMPY LMUSHKZ RYLMQRK XYBCLWU TMLYEYO RMTRYZM TOCYKZE
MPYWMOQ YZKZVYZ YZVYLPY RLMSHRY LBYTVYL MUOBUYT TYZQMZZ
VKYMUOA YLRUZEO CTSHYLW MOQYZKO RVMZZMU BEMPYXC ZOCEYZM
ZZRYZSE KOSLKFR OVYLYZF LCELMWW KYLUZEO KYKZRYK TKKKUZV
KXVKYOY OQULOYO TYLZYZQ CYZZYZM UBEMPYY KZOYKZY KZBMSHY
OBCLWUT MLVULSH VKYOYMU BEMPYTY LZYZOKY YKZBCLW UTMLJUV
YBKZKYL YZUZVXC ZYKZYWA AAOYLXY LMUOAYL RYZJUTM OOYZMUB
EMPYJAY KVKYYTY WYZRYXC ZBCLWUT MLYZBCL WUTMLYK ZHRWTQM
ZZWMZWK RYKZEMP YBYTVYL ZMQRKCZ OPURRCZ OMUOAMH TTKORYZ
UZVXKYT YWAMOHY URYOCZO RZCSHOC MZCPYLB TMYSHYZ YTYWYZR
YZUYPTK SHKORPY ORUYSQY ZVKYWCY ETKSHQY KRYZHKY LJUTYLZ
YZOKYKZ VKYOYLM UBEMPYQ YZZYZMU BEMPYVL YKEYORM TRUZEXC
ZYKZEMP YWMOQYZ KZVYZYL ORYZPYK VYZMUBE MPYZHMP YZOKYVK
YBCLWUT MLYTYWY ZRYYKZB MSHKLEY ZVACMUB VYLOYKR YHYLUWO
RYHYZTM OOYZACO KYEYLMV YJUBMYT TKEHKZE YLMRYZA MLYZZMR
UYLTKSH WCYSHRY ZOKYTKY PYLYKZY ZEYBMYT TKEYZPY ZURJYLB
LYUZVTK SHYZMUB PMUYKZY LYKZEMP YWMOQYY LLYKSHY ZAKYWMZ
VMOWMSH YZQMZZT YLZYZOK YHKYLMU BEMPYXK YLYKZYV MRYZPMZ
QWMOQYV MYOOKSH WKRYKZY WQCZQLY RYZJKYT XCLMUEY ZTYKSHR
YLTYLZR OCTTYZO KYHKYLY KZYYKZE MPYWMOQ YBUYLYK ZYMQRUY
TTYVMRY ZPMZQMZ AYZVUZE KWQTKZK QUWVYBK ZKYLYZU ZVEYORM
TRYZ

Breaking the Monoalphabetic Substitution

We try frequency analysis and make the following conjectures in the order from (1) to (6):

Most frequent letters:         Conjecture:

Y  213 18.8 %                  Y = e (1)
Z  137 12.1 %                  Z = n (2)
K   98  8.7 %                  K = i (3)
M   88  7.8 %
L   70  6.2 %
O   61  5.4 %
R   53  4.7 %
T   53  4.7 %
U   51  4.5 %
W   39  3.4 %
V   33  2.9 %
E   32  2.8 %
H   32  2.8 %
B   31  2.7 %
C   31  2.7 %
P   28  2.5 %
Q   25  2.2 %
S   20  1.8 %
A   17  1.5 %
X   10  0.9 %

The most frequent pairs, using (1) - (3) :

YZ   frequency:  57  en
YL   frequency:  41  e*    L = r (4)
KZ   frequency:  31  in
YK   frequency:  31  ei
KY   frequency:  29  ie
RY   frequency:  26  *e    R = t (6)
PY   frequency:  22  *e
ZY   frequency:  19  ne
LY   frequency:  18  *e
SH   frequency:  17        S = c, H = h (5)
MU   frequency:  17

Partially Decrypted Text:

AAAQUrO teiTiiB CrWUTMr einhtWT PiOherh MPenOie EeTernt
AieWMni nBCrWMt iCnOOei teninht WTVeBin iertWit htWTQMn
nWMnMPe rMUchin terMQti XeBCrWU TMreEeO tMTtenM TOCeinE
MPeWMOQ eninVen enVerPe trMchte rBeTVer MUOBUeT TenQMnn
VieMUOA ertUnEO CTcherW MOQeniO tVMnnMU BEMPeXC nOCEenM
nntencE iOcriFt OVerenF rCErMWW ierUnEO ieintei TiiiUnV
iXVieOe OQUrOeO TernenQ CennenM UBEMPee inOeine inBMche
OBCrWUT MrVUrch VieOeMU BEMPeTe rnenOie einBCrW UTMrJUV
eBinier enUnVXC neineWA AAOerXe rMUOAer tenJUTM OOenMUB
EMPeJAe iVieeTe WenteXC nBCrWUT MrenBCr WUTMrei nhtWTQM
nnWMnWi teinEMP eBeTVer nMQtiCn OPUttCn OMUOAMh TTiOten
UnVXieT eWAMOhe UteOCnO tnCchOC MnCPerB TMechen eTeWent
enUePTi chiOtPe OtUecQe nVieWCe ETichQe itenhie rJUTern
enOiein VieOerM UBEMPeQ ennenMU BEMPeVr eiEeOtM TtUnEXC
neinEMP eWMOQen inVener OtenPei VenMUBE MPenhMP enOieVi
eBCrWUT MreTeWe nteeinB MchirEe nVACMUB VerOeit eherUWO
tehenTM OOenACO ieEerMV eJUBMeT TiEhinE erMtenA MrennMt
UerTich WCechte nOieTie Pereine nEeBMeT TiEenPe nUtJerB
reUnVTi chenMUB PMUeine reinEMP eWMOQee rreiche nAieWMn
VMOWMch enQMnnT ernenOi ehierMU BEMPeXi ereineV MtenPMn
QWMOQeV MeOOich Witeine WQCnQre tenJieT XCrMUEe nTeicht
erTernt OCTTenO iehiere ineeinE MPeWMOQ eBUerei neMQtUe
TTeVMte nPMnQMn AenVUnE iWQTini QUWVeBi nierenU nVEeOtM
Tten

This looks good and hints at a German plaintext. For M we could substitute s or a; the already identifyed text snippets make »a« much more plausible because it is a vocal. For s now virtually only O makes sense. If we moreover substitute P by b, then in the first row the text »bisher haben sie« appears—a clear affirmation of the correct way. The remainder is routine and a little bit of trial and error.


Exercise. Perform the complete cryptoanalysis of the ciphertext whose period you determined in a former exercise.

At this point it may be interesting to follow the cryptanalysis of the cryptogram from Jules Verne's Jangada.

Note. That this simple adjustment of the columns worked is a fortunate coincidence. In the next section we'll examine this phenomen.


Author:
Klaus Pommerening, 1997-Jul-14; last change: 2014-Jan-16.