Challenge Eight - third clue

The text consists of 2320 characters and 2320 factorises as 5*16*29, so the keyword for the transposition phase is likely to have length 4,8 or 10 (other lengths are possible but unlikely). Let's just say that you might want to think like Goldilocks. When you are trying to pair the columns remember that you are looking for the roughest distribution. In particular of the 36 possibilities you really only expect the text to use around 26 of them (maybe one or two numerals and all or most of the letters) whereas if you pair the columns wrongly you are likely to see almost all 36 possibilities appear. Keep trying!

Challenge Eight - second clue

Had a request to say a bit more about step 2. think about what Aggie does. First she substitutes pairs of letters for individual letters. Next she writes the new text (which just uses six lettters used over and over again) in columns so the first two columns go together then the next two go together and so on . Then she shuffles the columns. When she does that she breaks up[ the pairs (if the first four characters are now CVSE the CV go together and the SE together before shuffling. Maybe after shuffling she gets CEVS splitting up the CV and the SE and switching the order of the S and the E). Step two is a description of how to see which columns pair up (in the example it would tell you to pair column one with column three and column two with column four). So for example you try pairing columns and and two and do a frequency analysis on the letter pairs that gives from those two columns. Next you pair one and three and do the frequency analysis for that pairing and so on. One of the pairings (one with something) will give you a much rougher distribution - fewer pairs will be used and one or two will be used much more than the others etc.. That one is the correct pairing for column one (in the example you would see ti was the pairing of column one with column three) Next you repeat to find out what to pair the next column with and so on. All the best,

Challenge Eight - first clue

This looks strange but it seems to be a reasonably standard ADFGVX cipher, using the characters S,U,E,V,I,C instead! Also, we know that Aggie has a habit of reading across the rows not down the columns in the transposition stage of the cipher. Your task is big.

Step 1. Guess the keyword length - another clue about this later
Step 2. Figure out how to transpose the columns. This is really tricky, but there is a way of cutting down the possibilities. The columns come in pairs, and if you can figure out which ones to pair together that cuts the possibilities a lot. To find out pair them all the ways you can and in each case do a frequency analysis to see what distribution of pairs you get. If column x pairs with column y then the frequency distribution will be rougher for that pairing than for any pairing x with another column. Start by trying to figure out which column pairs with column 1 then try to find a second pair from the other columns and so on. I used Excel to do this.
Step 3. Once you know how to pair the columns you need to work out for each pair which order they come in, so knowing that colummn x and column y are paired you need to know whether to write the pairs in the order xy or yx. To find out do a frequency analysis of occurences of the 36 letter pairs for each choice (xy and yx) for each of your pairs. You want to choose the ordering so that you get roughly the same frequencies for all the pairings.
Step 4. Once you have done this you have reduced the transposition stage from 2n columns of individual letters to n columns of letter pairs. There will be n! ways to rearrange the columns and each rearrangement can be viewed as a simple subsitution cipher. Of course only one of them will give you the answer, so you need to examine each one in turn and try frequency analysis. With some effort you will crack the cipher.

I will give more clues as time goes on.

Challenge Seven - solution

How did you get on? A small variant on the ADFGVX cipher slowed you right down, but 96 correct entries was pretty good. I won't say too much about the solution, 'cause I figure you're pretty focused on the final challenge. The fifty pounds cash prize winners this week are:

James Lloyd from Portsmouth Grammar School
Jan Gondzio from James Gillespie's High School
Morty's Marvels (that's Thomas Woolley, Jamie Stone, Shani Martin, Sean Renshaw) from Mortimer Wilson School
Timothy Gale from The Leventhorpe School
Booth Dilkes (Jim and Silas Dilkes) from Rodborough School.

Congratulations to all of you and good luck with Challenge 8.

Challenge Seven - second clue

Still stuck? Once you rearrange the columns using the keyword you need to work out the subsitution. It looks tricky since each plaintext character is substituted by a pair of characters, but it is still a substitution cipher. If it makes it easier, just replace each pair (AA, AD, AF etc.) by a sinlge symbol from the set A-Z together with the integers 0 to 9, then treat the result as a simple substitution cipher. Good luck,

Challenge Seven - first clue

How are you getting on? Aggie threw you a curve ball with this one. It IS an ADFGVX cipher, but the standard way of writing out the cipher text is to work up the columns whereas Aggie seems to have worked across the rows instead. I figure she was nervous to use a standard German cipher with Germans on her tail. Any case, that should help. More later,

Winners Challenge Six

Congratulations to all of you who tried Challenge six. It was pretty tough and you did very well. We are pleased to announce the prize winners for that Challenge. As usual each team listed here wins 50:

The L33t H4x0rz of Varndean College,Brighton
Jan Gondzio of James Gillespie's High School,Edinburgh
Timothy Palmer of Haybridge High School,Stourbridge
Lucy Payne of St. Anne's Convent School,Southampton
Karim M Ladhu of North London Tutorial College, London

Challenge Seven

Looks like we are coming to the end of this trail, Aunt Aggie has returned to Southampton, presumably to steal back the slab. Even after you translate the heiroglyphics back into English you still have a long way to go with this one, so I've done the translation for you. Click on the Journal to get the text version. By the way, the text begins AXGVF and finishes with DFAVX AA, just in case you can't see it all. Let us know if you have problems downloading it.

You can find out about the cipher Aunt Aggie has used, just look it up on the web or in the back of Simon Singh's book. It's a two stage cipher, and you have to decrypt it in two steps. Hell, why don't I make it easy for you and tell you what the keyword is for the first part of the decryption: BABYLON. After that you are on your own, at least for a while. More clues later.

Challenge Six Solution

Much trickier this time and I have to say I'm impressed. Aunt Aggie got a taste for things Egyptian, and untangling those heiroglyphics certainly slowed me down. Of course you did have a crib, the same glyphs were used all over the site, (if not exactly in the same way!) and that should have helped with converting them to English characters. Once that was done you had the problem of another transposition cipher. This time the ciphertext had 642 characters in it, and 642=2.3.107, so the keyword length was likely to have length 6. The next step is to regroup the letters as

BTINA?
TEURTE
?YMTAH
UOI?ER
.....

where the question mark denotes heiroglyphs not used onthe website and for which you might not have known the meaning. A few experiments might have suggested reversing the columns to get

?ANITB
ETRUET
HATMY?
RE?IOU

leaving you to fill in the blanks.

Sounds like Aggie is off to Southampton next, and who can blame her. Fame and fortune await if she can get the slab to the collector in New York.