WoR Chapter 84 code

[Aminar]

Another thought. Pattern's name is a string of numbers.  This might have something to do with Cryptics....

[Kurkistan]

Split up according by longest-possible palindrome at each point, starting at the beginning.

 

111-8-2-5-101-11-2-7-1-2-4-9-151-2-10101-11-4-1-0-2-151-171-121-0-111-2-171-3-44-8-3-111-0-7-151-4-2-5-414-3-4-1-0-9-161-4-9-1-4-9-3-4-121-22-5-4-10101-2-5-1-2-7-101-5-191-0-111-2-3-4-1-2-55-11-525-121-575-5-111-2-3-4-101-11-2-9-151-2-1-0-6-1-5-3-4

There are 25 unique numbers here. Could this be a substitution cypher then?

 

 

Very interesting.

 

Coincidentally, all those palindromes sum to relatively small numbers:

 

 

[3, 8, 2, 5, 2, 2, 2, 7, 1, 2, 4, 9, 7, 2, 3, 2, 4, 1, 0, 2, 7, 9, 4, 0, 3, 2, 9, 3, 8, 8, 3, 3, 0, 7, 7, 4, 2, 5, 9, 3, 4, 1, 0, 9, 8, 4, 9, 1, 4, 9, 3, 4, 4, 4, 5, 4, 3, 2, 5, 1, 2, 7, 2, 5, 11, 0, 3, 2, 3, 4, 1, 2, 10, 2, 12, 4, 17, 5, 3, 2, 3, 4, 2, 2, 2, 9, 7, 2, 1, 0, 6, 1, 5, 3, 4]

EDIT: On reflection, I don't think this is the right track. Recall that no 0 appears without a 1, so any analysis should probably consider the sequence "10" to be a single character.

Edited March 7, 2014 by Kurkistan

[briangri]

Split up according by longest-possible palindrome at each point, starting at the beginning.

111-8-2-5-101-11-2-7-1-2-4-9-151-2-10101-11-4-1-0-2-151-171-121-0-111-2-171-3-44-8-3-111-0-7-151-4-2-5-414-3-4-1-0-9-161-4-9-1-4-9-3-4-121-22-5-4-10101-2-5-1-2-7-101-5-191-0-111-2-3-4-1-2-55-11-525-121-575-5-111-2-3-4-101-11-2-9-151-2-1-0-6-1-5-3-4

There are 25 unique numbers here. Could this be a substitution cypher then?

 

 

I ran it through a subsitution solver, and the best it came up with:

 

"dies then we orgem how aegqcadeqlzil dangoes blowar for worlock somes went spade lowevhxcus delotherge waywslo"

Maybe would work better with a Sanderson-focused solver?

[HydrogenAlpha]

Quick question: Is it rare to get full and proper words at "failed" attempts ? 

[Aminar]

Quick question: Is it rare to get full and proper words at "failed" attempts ?

No. He ran several attempts and found something. I still think we should try with the Alethibet's 25 letters and see what happens...

[Satsuoni]

Split up according by longest-possible palindrome at each point, starting at the beginning.

111-8-2-5-101-11-2-7-1-2-4-9-151-2-10101-11-4-1-0-2-151-171-121-0-111-2-171-3-44-8-3-111-0-7-151-4-2-5-414-3-4-1-0-9-161-4-9-1-4-9-3-4-121-22-5-4-10101-2-5-1-2-7-101-5-191-0-111-2-3-4-1-2-55-11-525-121-575-5-111-2-3-4-101-11-2-9-151-2-1-0-6-1-5-3-4

There are 25 unique numbers here. Could this be a substitution cypher then?

 

Tried that out in Brandon-geared hillclimber, still no luck. It may not work for Alethi.. or as well as some others, actually, though.

[Kurkistan]

Be sure to remember that 10 should probably be a singular character.

Edited March 7, 2014 by Kurkistan

[Krenn]

I ran it through a subsitution solver, and the best it came up with:

 

"dies then we orgem how aegqcadeqlzil dangoes blowar for worlock somes went spade lowevhxcus delotherge waywslo"

Maybe would work better with a Sanderson-focused solver?

 

what substitution solver are you using? i haven't been able to find a really good one yet.

[briangri]

what substitution solver are you using? i haven't been able to find a really good one yet.

I used this one, but I'm not sure how good it is.

 

http://www.blisstonia.com/software/WebDecrypto/

[Pattern]

Split up according by longest-possible palindrome at each point, starting at the beginning.

111-8-2-5-101-11-2-7-1-2-4-9-151-2-10101-11-4-1-0-2-151-171-121-0-111-2-171-3-44-8-3-111-0-7-151-4-2-5-414-3-4-1-0-9-161-4-9-1-4-9-3-4-121-22-5-4-10101-2-5-1-2-7-101-5-191-0-111-2-3-4-1-2-55-11-525-121-575-5-111-2-3-4-101-11-2-9-151-2-1-0-6-1-5-3-4

There are 25 unique numbers here. Could this be a substitution cypher then?

 

 

At least the splitting is well defined. I also have tried substitution (and back we are), but didn't get something proper of it.

 

Concerning Glyphs: I am not quite sure if they can be "constructed" properly. There seems to be much freedom in shaping the

glyphs. You can see this with the "Alight, wind approach deadly approaching winds alight" ketek: "The glyphs of the ketek were drawn in the shape of two storms crashing into one another" - implying they could also have been written in another form, probably keeping the connections but stretching in one direction and squeezing in another (mapping via diffeomorphism). This morphability

can also be seen with the bridgemen tattoos. Glyphs seem to have been reshaped, so that the total tattoo is somewhat aestetic.

 

Edit: Harakeke is on a very promising approach to build glyphs...

 

Especially "Kholin" glyph and "Tanat" being nearly mirror images of each other disturbs me, since in WoK the "kholin" glyph has been described as glyphpair "khokh" + "linil" looking like a tower and a crown. "Tanat" would be a single glyph, since it is a palindrome. There is either a mistake or glyphs are all but fixed.

Edited March 10, 2014 by Pattern

[Harakeke]

At least the splitting is well defined. I also have tried substitution (and back we are), but didn't get something proper of it.

 

Concerning Glyphs: I am not quite sure if they can be "constructed" properly. There seems to be much freedom in shaping the

glyphs. You can see this with the "Alight, wind approach deadly approaching winds alight" ketek: "The glyphs of the ketek were drawn in the shape of two storms crashing into one another" - implying they could also have been written in another form, probably keeping the connections but stretching in one direction and squeezing in another (mapping via diffeomorphism). This morphability

can also be seen with the bridgemen tattoos. Glyphs seem to have been reshaped, so that the total tattoo is somewhat aestetic.

 

Especially "Kholin" glyph and "Tanat" being nearly mirror images of each other disturbs me, since in WoK the "kholin" glyph has been described as glyphpair "khokh" + "linil" looking like a tower and a crown. "Tanat" would be a single glyph, since it is a palindrome. There is either a mistake or glyphs are all but fixed.

From what I can tell, there seems to be a method to writing glyphs in which the first sub-glyph is written in the top center, and the other sub-glyphs nest down and out from there. So while there is a lot of freedom, there's a general pattern that they seem to follow.

 

I haven't worked through the Kholin and Tanat glyphs in the tattoo yet, but I do notice some very subtle differences between them. I think they were deliberately designed to look very similar.

[Pattern]

Thats what I wanted to say - the design seems to be very deliberate. Glyphs are more like pictures than script. Like Egyptian Hieroglyphs.

And still different. Assigning a number - therefore the code to glyphs looks to me quite difficult.

I fear we are drifting to another problem - decrypting glyphs, instead of solving the code...

[DefiantBurrito]

This looks fun, I'm trying to catch up to see if I can help in some way...

 

IMHO:

 - I don't see a really strong connection to glyphs yet, might be a dead end.

 - I don't think the "treat every 0 as 10" is necessarily valid... it seems like there are enough 1's in there that it could have easily happened by chance.

 - I think the massive palindrome/ketek is the most significant discovery so far, because it seems REALLY unlikely to me that it could happen by chance!  I'm going to think about that a bit more and see what I can come up with...

[Nikita]

Guess my previous suggestion is out.

 

Looking at the digit frequencies, combined with 0 always following 1, I can't shake off the sense that this has something to do with two-digit numbers, with lots in the 10's (and maybe 20's)

1 - 66
2 - 20
5 - 17
4 - 15
0 - 13
3 - 7
7 - 6
9 - 6
6 - 2
8 - 2

[taveren]

come on guys you can do it break that code

Edited March 7, 2014 by taveren

[Galladon]

 

 

Guess my previous suggestion is out.

 

Looking at the digit frequencies, combined with 0 always following 1, I can't shake off the sense that this has something to do with two-digit numbers, with lots in the 10's (and maybe 20's)

1 - 66
2 - 20
5 - 17
4 - 15
0 - 13
3 - 7
7 - 6
9 - 6
6 - 2
8 - 2

this just seems to me like the classical Benfords's law- information is stored (in non randomized sytems) not linear but loagrithmic this means in any given text (e.g. wallstreet journal) you'll find numberes in the following distribution:

1:0.3

2:0.18

3:0.13

4:0.10

5:0.08

6:0.07

7:0.06

8:0.05

9:0.04

 

[Krenn]

I found a free code solving tool, here:  I haven't tested it much, but hopefully it will help us solve this.
I suspect the next step will be upgrading the program's dictionary to be compatible with fantasy terms from Words of Radiance

https://sites.google.com/site/cryptocrackprogram/

 

the program gives these code types as the most likely, with the lowest score being more likely:


Nihilist Sub...........33
Monome Dinome..........49
Tridigital.............52
Double CheckerBoard....72
Grandpre..............101
Morbit................102
Patristocrat..........120
Bazeries..............138
Route Transp..........150

Edited March 7, 2014 by Krenn

[Aminar]

This looks fun, I'm trying to catch up to see if I can help in some way...

 

IMHO:

 - I don't see a really strong connection to glyphs yet, might be a dead end.

 - I don't think the "treat every 0 as 10" is necessarily valid... it seems like there are enough 1's in there that it could have easily happened by chance.

 - I think the massive palindrome/ketek is the most significant discovery so far, because it seems REALLY unlikely to me that it could happen by chance!  I'm going to think about that a bit more and see what I can come up with...

Glyph's seem to be like making a picture out of letters. Like symmetrical Ascii art that says something. But the glyph's can be warped to make them.

[Brybry]

Something interesting I have found by dumping the ebook to HTML and then looking at it with a hex editor is that there is a UTF-8 zero width space (hex E2 80 8B) character after every 3 digits (though this leaves a single trailing 4)

I also noticed it in a lot of places where people have copy/pasted the number so I don't think it was just me.

 

I mention this because the only other place zero width spaces are used in the entire book is after every word in the two floorboard 17 every-second-letter epigraphs. Maybe Sanderson copy/pastes from his own encoder/decoder...

[Aminar]

That's an interesting find... I'm not sure what to make out of it... How many three digit combinations are there?

Edited March 7, 2014 by Aminar

[Fabrial]

Nice find Brybry. Might we use an html copy of the ebook to search for more keteks of the form "a b c X c b a" using Python/Perl? I'd help but I only have a hardcopy of the book

 

Maybe then we could find one that matched our palindrome number sequence in some way.

Edited March 7, 2014 by Fabrial

[Brybry]

http://pastebin.com/W0FvJ5xt

 

I'm not entirely sure it means anything and I'm not really sure the ebook is reliable. For one there are no line breaks -- it's just one big number. There are also two separate spans with the number (and the same zero width spaces) but the second is invisible alt text with occasional normal spaces.

Also, when I looked at it further the zero width spaces are sometimes in the middle of a word which is odd but it's still such a weird coincidence. Also, I would think that if any epigraph was going to be related that it would be the other Ceiling Rotation/Pattern epigraph.

[Lomeon]

That's an interesting find... I'm not sure what to make out of it... How many three digit combinations are there?

 

There are 37 unique three-digit strings broken up like this:

 

111 825 101 112 712 491 512 101 011 141 021 511 711 210 111 217 134 483 111 071 514 254 143 410 916 149 149 341 212 254 101 012 512 710 151 910 111 234 125 511 525 121 575 511 123 410 111 291 512 106 1534

 

Here they are with how many there are out of how many times that three-digit sequence shows up in the unbroken string.

 

111x5/8

825x1/1

101x3/7

112x1/6

712x1/1

491x1/2

512x3/4

011x1/5

141x1/1

021x1/1

511x3/3

711x1/1

210x1/3

217x1/1

134x1/1

483x1/1

071x1/1

514x1/1

254x2/2

143x1/1

410x2/4

916x1/1

149x2/2

341x1/4

212x1/1

012x1/1

710x1/1

151x1/5

910x1/1

234x1/2

125x1/2

525x1/1

121x1/6

575x1/1

123x1/2

291x1/1

106x1/1

1534

 

Also, does the eBook have uneven line breaks like the hardback?

[Aminar]

There are 37 unique three-digit strings broken up like this:

 

111 825 101 112 712 491 512 101 011 141 021 511 711 210 111 217 134 483 111 071 514 254 143 410 916 149 149 341 212 254 101 012 512 710 151 910 111 234 125 511 525 121 575 511 123 410 111 291 512 106 1534

 

Here they are with how many there are out of how many times that three-digit sequence shows up in the unbroken string.

 

111x5/8

825x1/1

101x3/7

112x1/6

712x1/1

491x1/2

512x3/4

011x1/5

141x1/1

021x1/1

511x3/3

711x1/1

210x1/3

217x1/1

134x1/1

483x1/1

071x1/1

514x1/1

254x2/2

143x1/1

410x2/4

916x1/1

149x2/2

341x1/4

212x1/1

012x1/1

710x1/1

151x1/5

910x1/1

234x1/2

125x1/2

525x1/1

121x1/6

575x1/1

123x1/2

291x1/1

106x1/1

1534

 

Also, does the eBook have uneven line breaks like the hardback?

Peter has said the line breaks are irrellevent. But yes it does.

And that's 30 some deviations. That's too many for capitol letters unless there are a lot of names, so I don't think it works.

[PeterAhlstrom]

The zero-width spaces in the ebook are there just to make the paragraph reflow without the ebook reader adding hyphens. They and their placement don't have anything to do with the code.