[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

VMs: A cryptological assault on Strong's decryption...

Hi everyone,

At 05:03 13/08/2004 -0600, GC wrote: 
I didn't say Strong based his decipherment on
Feely's decipherment, only that he took 'cribs' from the publications of
Feely and O'Neill.  I don't need to repeat what they were, they're published
in Strong's letters, even the discussion of Feely's decipherment and
Strong's opinion on the matter.  I think Strong's meeting with O'Neill had a
much larger influence than Feely's

Just to save everyone else the trouble, I thought I'd go through Strong's 111 letters (again), and extract the (few) places he discusses his system (appended below). I didn't notice any mention of any specific cribs anywhere - GC, please correct me if I'm wrong, or if you're actually referring to his 1945 article here?

For the record, I have tried to replicate Strong's decoding method (as described by GC, but without the half-spaces tweak) though in a more overtly cryptological way, but without any success. While GC is a very much more accomplished code-breaker than I am, I must confess that I did expect a ~little~ more success in that endeavour than ~none~ - and so came away believing that Strong had probably read more into those few pages he had available than was actually there.

Perhaps GC's new half-spaces tweak would be sufficient to get a very much better result... if that's the only difference, then I should be able to give it another go without vast difficulty. Just so everyone can follow what I did (and would do again), here are the details:-

Cryptologically: decomposing the magic (sorry, astronomical) key 135797531474 yields a 12-cycle sequence of six multiple alphabets, where the weakest link in the system is alphabet 7 (which appears three times in the sequence, whereas alphabet 9 appears only once), which would hence make up 25% of the words in the text. FWIW, ISTM that a sequence of 1357975314*9*4 would be a little more secure, but that's magic for you. :-o

  1-------1---  // +0, +8               // gaps of +8 and +4
  -3-----3----  // +1, +7               // gaps of +6 and +6
  --5---5-----  // +2, +6               // gaps of +4 and +8
  ---7-7----7-  // +3, +5, +10  // gaps of +2, +5, and +5
  ----9-------  // +4                   // gap of +12
  ---------4-4  // +9, +11              // gaps of +2 and +10

If the 12-cycle sequence clicks forward [ie selects the next alphabet along in the magic key] every space (or even every half-space), then - even if you don't know the alphabet, or even the transcription - it should be possible (as long as you have the spaces/half-spaces correctly notated) to test each of the 12 positions in the cycle against the 7-pattern to see which position's distributional profile is furthest away from the overall statistics. That is, something like:-

        for (offset = 0 .. 12-1)
                reset counts
                for (pos = 0 .. paragraph_length-1 step 12)
                        add word # (pos + (offset +  3)  modulo 12) to count
                        add word # (pos + (offset +  5)  modulo 12) to count
                        add word # (pos + (offset + 10)  modulo 12) to count

This should yield a distinctive distribution to the stats, with a strong peak for the correct position in the cycle (for alphabet 7), but also with a second subpeak for alphabet 4 (ie an offset of six in the cycle from the best offset found), with all other results being far closer to the null distribution. This is because only alphabet 4 shares a +2 gap between its offsets with alphabet 7 (see the gap values in my table above), and hence should give a better than null result (but not as good as the alphabet 7 result).

Sorry for not posting this to the list before, but I didn't think anyone bar me was particularly interested in attempting a proper cryptological assault on Strong's method.

In my own tests, I didn't find any results which suggested that any such position existed with any statistical significance (I looked specifically at each of the two paragraphs on f78r, which Strong had been looking at, but not with the suggested half-space tweak), but perhaps someone else will have more luck with a half-space transcription, or from looking at different pages. Or perhaps I misunderstood GC's exposition of the whole system - but it certainly wasn't from the lack of trying he accuses me of.

Elmar, is this sufficient detail for you?

Cheers, .....Nick Pelling.....

* * * * * *

http://www.ehabitat.demon.co.uk/vms/letter-064.html
        The cipher is apparently a string cipher
        beginning at the first and continuing to the end.

http://www.ehabitat.demon.co.uk/vms/letter-091.html
             The key to the code can be found in Porta and Trithemius.
        It is also discussed in the Cryptomenes of Gustavus Selenus, a
        copy of which I own.  This is all I can say on the method.

http://www.ehabitat.demon.co.uk/vms/letter-094.html
        "There is a gradual approach to the distribution obtained by
        passing backward through the classical writers of the 17th and
        16th centuries in English.  The nearest approach was found to be
        in some early writing of John Dee at it was this characteristic
        which originally led me to believe that John Dee had something to
        do with its construction."..."If you will look on the last page
        of John Dee's Actions with Spirits which was published in 1659
        you will find a very interesting page.  By dividing the page into
        two parts you will find that one half has exactly the same
        frequency distribution of word size that is found in the Voynich
        manuscript.  If anybody be interested in working out this cypher,
        you will find the identical cypher which I determined in the
        Voynich manuscript.  The only difference being in the number of
        alphabets used and the difference in the arithmetical
        progressions."

http://www.ehabitat.demon.co.uk/vms/letter-097.html
             I admit that I do not have the complete multiple alphabet,
        but I have enough to convince me that I am on the right track.

http://www.ehabitat.demon.co.uk/vms/letter-102.html
             It was even suggested by one of our top cryptographers in
        Washington that since they had the mechanical machine to do the
        entire decipherment all they needed was my system of decoding
        which I had obtained by hard work.  This I was loath to do and so
        the matter rests.

http://www.ehabitat.demon.co.uk/vms/letter-106.html (from Jim Gillogly, 1975)
* I would be interested to see your method; I am particularly
* mystified that a progressive multiple alphabet system could
* produce situations like the three contiguous repetitions of the
* work 40HC89 in the plate on p. 865 of The Codebreakers.

http://www.ehabitat.demon.co.uk/vms/letter-107.html
             If you continue p. 865 you may find more repetitions of what
        you call "the word 40HC89".  You may conclude that this
        combination is not even a single word and the combination may
        have different meanings.  In other words it is not a unit of
        words but a unit for the entire passage.  That is as far as I
        will go in indicating my system.
             My description of the system of decoding used is clear.  But
        even without this clear hint, any expert decoder should be a
        master in his own field.

http://www.ehabitat.demon.co.uk/vms/letter-110.html
I shall never divulge the system
i used. I carefully described it in the Science article. you
can find the cypher in Trithemius. 


______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list