e 11e• 1 lll'11VOWW l1 IJf l JWOV fijlJllJf IrCDlilii' UllsllJlilUll Ul W13 i l ll3 9 00ullil l1£ii llilD EO 3 3b 3 PL 86-36 50 USC 3605 Non - Responsive Im I •t• BIUllh 118¢11 lH ill £ I f 1 Die Ea 11 u Ga i C k A fidtu1 I 8 Declassified and Approved for Release by NSA on 03-01-2021 pursuant to E O 13526 MDR-10935 9- EO 3 3b 3 PL 86-36 50 USC 3605 I - · tiV'-------- _ _ _ _ _ _ _ _ _ __ DEREK ·K CRAIG ________________ J Sep 74 • CRYPTOLOG • Page 1 EO 3 3b 3 PL 86-36 50 use T8PHGl FIM1IM 3605 EO 3 3b 3 EO 3 3b 6 PL 8 6- 36 50 USC 3 605 EO 3 3b 3 EO 3 3b 6 PL 8 6-36 50 US C 3605 ' ' Sep 74 • CRYPTOLOG Page 2 J8' lff Rff tMIIM EO 3 3b 3 EO 3 3b 6 PL 86-36 50 USC 3605 5ep 74 • CRYPTOLOG • Page 3 19' Miff IMIIM EO 3 3b 3 EO 3 3b 6 PL 86- 36 50 USC 3605 e J ell 'F WM Sep 74 Ci saca ll• l'jlt CRYPTOLOG • Pase 4 JQII lialf IMIIIM EO 3 3b 3 PL 86-36 50 USC 3605 ·-· • •• - -· - - ' • ·v- EO 3 3b 3 PL 86-36 50 USC 3605 Haff J l RYPTANALYSIS' ODE RECOVERY by ARJORIE MouNTJOY Ths rote of tM az yptanaZH•t in aods NeOnlltJouation ow r s that of ths Cl' IPUltinguiet and both ai-e ofttln odNNd bl tM aml l tum ''boolcbreaJc - If w aat¥Pt th rt1fin t pl'C1JH e d by Stuart Bllok Naogniud Agt111c1J authority on aods Naonet ruotion the W11Ptantit 11et bl'4141u inui a aod8 1 11site the oryptoZingwi •t gOfle on from thar- to ''J'aoonstJ MDt n tha cod book li't1 eOIINI noue pNpaNd esv11ral '1 _llaN Q O by Naiojol'i• l olmtJO' J Pl C'l'lfl'tanl2lyst l'IOIJ Nt Nd about what t 171'11Ptanat 11st in svafi a oass ehouZd 1cnol l ha _ or do to •MUN that t Ju 1 1hou pl'Or»H of odtls Hoonstruction Dl21I buit d a NZiabu bod I ol plll'tiMPft 1111Zttll'ia1 '1eN i n tll rrrlnDl' aZUl' Zt lona ZN th JlountiOII nottl8 TM11 dsmcmstrat8 Qgain on -v -- that age is not a true 1118aaure of r a't ld 'tl CfM cryPtanalyst and the code reconstructo shoi 1 wrk closely to1ether They uy even · be the sue person but in any case they should aim at c011patibility wherever possible 'Jbe co J tibility if especially desirable when it comes to lo11ing aessages and requesting machine runs 01' t alue ¥ Cods N_ ffl lOtoP• ehout d al ways hal18 t1aa11 accsae to th es 1111Ztli r€ale s rp ciat'l 11 1'1MN A tidy organization will desire the fol•auoa uor oodas l 1l'tl oonc msd TM etoN of lowing t_ools I say desire '' for they probably 111aul'iat1 and infomaticn is t JtOt for will not be pel'llitted to store so lar1e a th41 •aks of hoal'ding but to Hduoe the tw collection • rr1quirsd to s r ptoit a - In tlsi a buai• nothing ie •o 1t1t1 sufl l cu oonetcznt N • disaov Pfl of tu whesl •11••• The followin definitions and for11Ulae are of ut110st importance True bue 111• values assi ned are known to bathe sw as those used by the original cryptographers Relative bue The values assiped have the sw delta relationship as those of the original roups i e they are off by a constant of 1111 2222 etc Arbitra5 base The values assi ned can be convert to the true by the addition of a constant uiount which is purely arbitrary The Basic Cryptolo ic Glossary does not uke this distinction between relative and arbitrary base Jt should · Sep 74 • CRYPTOLOG • IEaEf Paae S 11111 11 St IICJ IIHlll I • II EO 3 3b 3 PL 86-36 50 USC 3605 l GINif ·-------------- n E Repeat rate of a code f1 f1-l • i•l I I N N-1 n Delta I C a • c • E i•l f 1 frl • N N-1 The followina relationships will be extremely useful to the analyst working The aaterialowas rec a • an rlllxaain • The - noB-dina• e hypothesis was duic'kly established but not before the crmfolinguist had sOJ1e aaonidng 111 Mnts • • ••• · ' 'l -- ' _____ J 'On the Before turning the material over to the cryptoltnauist for reconstruction of the codebook the cryptanalyst is responsible for ma'kina sure if possible ot er an t e c ryptana yst aist ave tJi e uraae of hls convictions and be willin1 to defend hts solution of a transposition bu st which yi1lds no repetitions at all • I den't want to get into the details 'of additi l'e solution but the wise cryptana yst will u e statistical infonation to autd him 2 that the aaterial is in fact plain code to tha shortest road to recovery The repeat rate of the code discernible fro• the d9Pths shou d tell hia whether the solution can•be 3 that the encipher aent has been c011plet ly obtained in a few days by hand Mthods ol and correctly re110Ved whefher it uy take a year or mre and a fullsceJ e 1119Chani ed attack to complete his p rt 4 that the aaterial is ho ageneous of•the code solution He should be able to 5 that the code is reduced to true base not spbt a one-part code fr011 the repeat rate of ti• be1inning digraphs always possible 1 that the code is not already recovered The five points uy see• obvious but 1 can cite some heartbreak in examples as well as so111e heartenina ones code reconstructor is oft de- Sep 74 • CRY I OG • Paae 6 IHIR' • ff • INIIIM • lilt IHIIIII - · --- - --- - ---· ______________________ EO 3 3b 3 PL 8 6-36 50 USC 3605 ' --- ······------------ i l 11 a 1-ti m to g i 1eN dJ' IP o 1- 1 a 11epm-at• Coda Lib1'0 Z' I· The LibN ey is tluig wd l dl a NHaNh omiter foio oodA a it is tlul HtoZ'lf fof' obaoi ta QM aurp1 ws cod6books f'l4J8 IJJ'ld othef' cods mauriala ta 1 ·nt for statis_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ p'tao• forNpOl'ts genAINZ o t u11F recontias alld __ _ _ n - - acde gathet'ing Code reconstructor and cryptanalyst uy not see eye to eye on machine work particularl ' on inverse frequency lists The fol'lller often requests that the low-frequency groups be dropped 8Jld the total gets lost That h an unfortunate loss for the cryptanalyst whoso best estimate or the probability of occunence of a high-frequency group is Th• cryptanalyst atl'Motio i maohiM ai ppo 't foi- bookbf'eaking and t '1 state f the art it ia a wtispt'ing of ol'iginat atudl•• and nriovativ• pl'OOedurss And it iB poBBib1-e to have oowp re1umaiwi IIl'ahives without bursting at the 11eama thanlu t ' miarofilm QM miorofia '1 f81 8M 'f i1t 1eee N 'I' needs such infotution to prepare log weights and to assnble lists of aroups with a specifi probability Ho also needs it to compute the repeat rate Note This problem is solved by the current practice of showina the relative frequency as well as tho absolute frequency of a code aroup Ed Both the code roconstructor and the cryptanalyst rely on statistical aethods thouah the former sees them as only a saatl part of hh store of tools He is lookin fDr a one-to-one correspondence between code aroups and words or phrases and he is autded by the frequencies of these itns The • easure of the extent of · that pidanco varies according ta the individual's role and point of view--cryptanalyst or cryptolinauist a The cryptanalyst is in 110re desperate need of statistical auidance He is often in the position of a man lookin for a needle in a haystack when the needle looks exactly like a piece of hay a11zo Buak'• OOl ffl8nts HZ uanoau the dtff111' in aaBHlllfltmt of the rot• of stat istios tn rHXU Noonetzowotion B• lflllill Cods NooruJtiowation 11 pz l maril l a 1 iguistic pl'OCIIB8 but i t inllOltlllB a Zct moN than ' Zanguage k s° 'tedge The oodll NQ 718tl'UOto1' is in fa ot an anal yet in tM bJ'Otldsst sB111111 of the Mn'd Bil•ids• kmguag11 fZu•n afl• ha '8t have an intimatll tedg11 of t '1 aountey undllr atud J its hist 'ey '41 tuN goo11f'l'ltlMmtal atJ'MotuN k11y P4JNonaliti '1 g11ograph¾J--a id a host of othar matiilll'lt Codeb oks ZN NoonetJoucted through knowtcdge not tmoough ignomnae 2 111 aim of coda Noonstruation is UJ 1111t4b1 i h r 1hat is pf'Otlabl•-not r ihat ie plawlible I Non - Responsive