Pattern he/him Posted March 4, 2014 Report Share Posted March 4, 2014 AhbuttheywereleftbehindItisobviousfromthenatureofthebondButwherewherewhereSetoffObviousRealizationlikeapricity TheyarewiththeShinWemustfindoneCanwemaketouseaTruthlessCanwecraftaweapon - From the Diagram, Floorboard 17: paragraph 2, every second letter starting with the first TherehastobeananswerWhatistheanswerStopTheParshendiOneofthemYestheyarethemissingpiecePushfortheAlethi todestroythemoutrightbeforethisoneobtainstheirpowerItwillformabridge - From the Diagram, Floorboard 17: Paragraph 2, every second letter starting with the second These lines would have been mingled in the Diagram like AThhbeurtethhaesytwoebreealneafnts....looks very gibberish, but isn't. Perhaps this has to be applied to decoded 11182510111271249151210101114102151171121011121713 44831110715142541434109161491493412122541010125 1271015191011123412551152512157551112341011129151 21061534 2 Link to comment Share on other sites More sharing options...
Aminar Posted March 4, 2014 Report Share Posted March 4, 2014 1 64 2 18 3 7 4 14 5 13 6 2 7 5 8 2 9 6 0 12 Of note, every 0 is preceded by a 1, so that makes 52 1s. This dispersion is strange, and not random in the least. Link to comment Share on other sites More sharing options...
Pattern he/him Posted March 4, 2014 Report Share Posted March 4, 2014 (edited) Thanks for counting the numbers! This distribution hints to substituting doublets with letters. Using latin alphabet (I don't know Alethi) we would get a "natural" table 1 2 3 4 5 6 7 8 9 10 A B C D E F G H I J11 12 13 14 15 16 17 18 19 20 K L M N O P Q R S T21 22 23 24 25 26 U V W X Y Z starting with 1 and ending with 26 or 0 1 2 3 4 5 6 7 8 9 A B C D E F G H I J10 11 12 13 14 15 16 17 18 19 K L M N O P Q R S T20 21 22 23 24 25 U V W X Y Z starting with 0 and ending with 25. These still can be permutated. But what you see is, that 1 should be abundant, since it appears with 12 letters (of 26), while the other numbers are much rarer. The fact that a 0 is always preceded by a 1 hints to a numbering from 1 to 26 and not from 0 to 25, although one cannot be sure about it. If we take the doublet 10 then, it appeares 12 times, which is not too often for an abundant letter like "e" or "t". Then there are my favourites 111 - are they 1 11 or 11 1 or something else entirely (embedded in numbers)? Another ambiguity: 25 for example can be "25" or "2" - "5" Trying to keep all 10s, I try decomposition: 1 11 8 25 10 1 11 2 7 12 4 9 15 12 10 10 1 11 4 10 21 5 11 7 11 2 10 11 12 17 134 4 8 3 11 10 7 15 14 25 4 14 3 4 10 9 16 14 9 14 9 3 4 12 12 25 4 10 10 12 512 7 10 15 19 10 1 11 23 4 12 5 5 11 5 2 5 12 15 7 5 5 11 1 23 4 10 11 12 9 1512 10 6 15 3 4 Mind, i split some numbers up arbitrarily - will have to puzzle a bit...I have found a nice tool (but still no result): http://cryptoclub.org/tools/cracksub_topframe.php Edited March 4, 2014 by Pattern Link to comment Share on other sites More sharing options...
Aminar Posted March 4, 2014 Report Share Posted March 4, 2014 (edited) Remember that Althis only has 25 letters. Which could mean 1-25. The high number of 111's implies some form of frequent letter combination... We need to get the Alethi Font guy in here... Also 5 is a double letter. Several 55s show up... 14 9 shows up twice in a row? 11 18 25 10 1 1 12 7 1 24 9 15 12 10 10 11 14 10 21 5 1 17 1 12 10 11 12 17 13 4 4 8 3 11 10 7 15 14 25 4 14 3 4 10 9 16 14 9 14 9 3 4 12 12 25 4 10 10 12 5 12 7 10 15 19 10 11 1 23 4 12 5 5 1 15 25 12 15 7 5 5 11 1 23 4 10 11 12 9 15 12 10 6 15 3 4 Here's a possible doublet set to work with... And Here's the Alethibet. A B D E F G H I J K L M N O P R S T U V Y Z SH TH CH I don't think it's a straight substitution. We need to work out the vowels. Edited March 4, 2014 by Aminar 1 Link to comment Share on other sites More sharing options...
Harakeke Posted March 4, 2014 Report Share Posted March 4, 2014 (edited) I've been playing around with the way Pattern split up the numbers into 1-2 digits in post #13, trying to find keys in English and Alethi, but nothing's panned out. Here's the spreadsheet I've been using, if that's helpful to anyone: https://drive.google.com/file/d/0B19mpDUN_8rSYjhsZGhlNFQwN1U/edit?usp=sharing Edited March 4, 2014 by harakeke 1 Link to comment Share on other sites More sharing options...
Pattern he/him Posted March 4, 2014 Report Share Posted March 4, 2014 (edited) I keep on working with Latin alphabet, the rest of the epigraphs is written in it (there are c's). Otherwise, the Diagram is not written in Alethi but in an invented language,which is more accurate (see T's interlude). To go there would render it unsolvable. Combination 26 does not occur, diphtongs would be a problem with existing tools. Edited March 4, 2014 by Pattern Link to comment Share on other sites More sharing options...
HydrogenAlpha he/him Posted March 4, 2014 Report Share Posted March 4, 2014 (edited) Big thumbs up to every one of you trying to solve this code ! I could never come up with something like this.Just a idea but maybe 111 indicates a capital letter ? Edited March 4, 2014 by HydrogenAlpha Link to comment Share on other sites More sharing options...
Pattern he/him Posted March 4, 2014 Report Share Posted March 4, 2014 Ok, it gets late - I will continue puzzeling tomorrow, as long as it is not solved until then. Good luck!! Link to comment Share on other sites More sharing options...
Harakeke Posted March 4, 2014 Report Share Posted March 4, 2014 (edited) I keep on working with Latin alphabet, the rest of the epigraphs is written in it (there are c's). Otherwise, the Diagram is not written in Alethi but in an invented language,which is more accurate (see T's interlude). To go there would render it unsolvable. Combination 26 does not occur, diphtongs would be a problem with existing tools. That's good to know. I'm still waiting for Amazon to deliver my copy, so I'm just going by the information in this thread for now. I agree that the Alethi women's script probably isn't the way to go. I'll play around more using the Latin alphabet when I get a chance. I like your theory that it's a two-level substitution/transposition cipher. (I tried brute-forcing various substitution solutions this morning, but nothing popped up.) In that case, what we'll need to do to crack this is: 1. Determine the proper parsing of the original number string 2. Determine the substitution key for translating the cipher text 3. Determine the ordering of the transposition cipher (likely solved) Edited March 4, 2014 by harakeke Link to comment Share on other sites More sharing options...
Voidbringer Posted March 4, 2014 Report Share Posted March 4, 2014 I'm not sure if this has been mentioned so here goes nothing. The Alethi Calendar has 10 months, 10 weeks per month, and 5 days per week. The year in which Bridge Four rescues Dalinar and Talenel returns is 1173. This corresponds to the first 4 digits of the following dates. The next six digits appear to be the the month, week (perhaps reversed), followed by the day of the week. 1173090605 1173090801 1173090901 1173091001 1173091004 1173100105 1173100205 1173100401 1173100603 1173100804 - From the Diagram, North Wall Coda, Windowsill region: paragraph 2 (This appears to be a sequence of dates, but their relevance is as yet unknown) How this applies to the code itself i surely don't see yet. I hope this wasn't too obvious Link to comment Share on other sites More sharing options...
Voidbringer Posted March 4, 2014 Report Share Posted March 4, 2014 (edited) The fact that there is no punctuation in what appears to be the translated versions of codes (a few posts above) at least indicates that there is no need to look for extra symbols. This might indicate it is just a substitution cipher (I can't imagine Brandon would get too complex). Edit: Oops. Double post. Edited March 4, 2014 by Voidbringer Link to comment Share on other sites More sharing options...
Harakeke Posted March 5, 2014 Report Share Posted March 5, 2014 (edited) I'm not sure if this has been mentioned so here goes nothing. The Alethi Calendar has 10 months, 10 weeks per month, and 5 days per week. The year in which Bridge Four rescues Dalinar and Talenel returns is 1173. This corresponds to the first 4 digits of the following dates. The next six digits appear to be the the month, week (perhaps reversed), followed by the day of the week. 1173090605 1173090801 1173090901 1173091001 1173091004 1173100105 1173100205 1173100401 1173100603 1173100804 - From the Diagram, North Wall Coda, Windowsill region: paragraph 2 (This appears to be a sequence of dates, but their relevance is as yet unknown) How this applies to the code itself i surely don't see yet. I hope this wasn't too obvious AhoyMatey speculated earlier that they might be the dates of highstorms, but I just calculated the intervals between the dates (56, 50, 50, 3, 51, 50, 96, 102, 101), and they seem to spread out for that, since highstorms come every few days. They remind me more of the dates for the deathrattle quotes in WoK. edit: Ah. The last death quote in WoK is dated Tanatanev 1173, which would be 1173090204. The list from WoR picks up not long after that with 1173090605 (Tanatashah? 1173). edit2: And just for the heck of it, here are the WoK dates rendered in North Wall Coda format: 1173010304, 1173030202, 1173040204, 1173040503, 1173050101, 1173050402, 1173050804, 1173060202, 1173060704, 1173070201, 1173070202, 1173080204, 1173080405, 1173080605, 1173080801, 1173080903, 1173090103, 1173090201, 1173090204 edit3: counting is hard =P Edited March 5, 2014 by harakeke 6 Link to comment Share on other sites More sharing options...
HydrogenAlpha he/him Posted March 5, 2014 Report Share Posted March 5, 2014 If you take a look at - for example - this: AhbuttheywereleftbehindItisobviousfromthenatureofthebondButwherewherewhereSetoffObviousRealizationlikeapricity TheyarewiththeShinWemustfindoneCanwemaketouseaTruthlessCanwecraftaweapon - From the Diagram, Floorboard 17: paragraph 2, every second letter starting with the first There a obviously not spaces, periods, and other punctuation marks in those sentences, but there are capital letters. I imagine that the code has to have those capital letters, too. Maybe the frequently occuring 111's indicate a capital Letter. 1 Link to comment Share on other sites More sharing options...
cris34b he/him Posted March 5, 2014 Report Share Posted March 5, 2014 Year 1173, Month 09, Week 06, Day 05. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 09, Week 08, Day 01. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 09, Week 09, Day 01. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 09, Week 10, Day 01. 2 days between dates. (not including the actual dates written here.) Year 1173, Month 09, Week 10, Day 04. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 01, Day 05. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 02, Day 05. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 04, Day 01. 6 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 06, Day 03. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 08, Day 04. This looks very much like a list of high storm dates, and the regularity of their occurrences, all but 2 happening 5 days apart from each other lends to the theory. Link to comment Share on other sites More sharing options...
Harakeke Posted March 5, 2014 Report Share Posted March 5, 2014 (edited) Year 1173, Month 09, Week 06, Day 05. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 09, Week 08, Day 01. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 09, Week 09, Day 01. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 09, Week 10, Day 01. 2 days between dates. (not including the actual dates written here.) Year 1173, Month 09, Week 10, Day 04. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 01, Day 05. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 02, Day 05. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 04, Day 01. 6 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 06, Day 03. 5 days between dates. (not including the actual dates written here.) Year 1173, Month 10, Week 08, Day 04. This looks very much like a list of high storm dates, and the regularity of their occurrences, all but 2 happening 5 days apart from each other lends to the theory. Righhhht - 5 days per week, not 50. *facepalm* Okay, new theory - death quotes correspond with highstorm occurrences! Though I did notice something while playing around with the dates - numbers <10 are always preceded by a 0 in the concatenated form. I'll play around with parsing the cipher numbers with that in mind. edit: 11 18 25 10 11 12 71 24 91 51 21 01 01 11 41 02 15 11 71 12 10 11 12 17 13 44 83 11 10 71 51 42 54 14 34 10 91 61 49 14 93 41 21 22 54 10 10 12 51 27 10 15 19 10 11 12 34 12 55 11 52 51 21 57 55 11 12 34 10 11 12 91 51 21 06 15 34 Maybe numbers >26 represent capitals? Edited March 5, 2014 by harakeke 1 Link to comment Share on other sites More sharing options...
cris34b he/him Posted March 5, 2014 Report Share Posted March 5, 2014 Just realized the dates I posted were not the actual code, but instead were example dates. *facepalm* And back from scratch we go... 1 Link to comment Share on other sites More sharing options...
Satsuoni he/him Posted March 5, 2014 Author Report Share Posted March 5, 2014 Well, using 10 as a code didn't end up with any proper splits as all. And transposition -or simply mixing the letters invalidate statistical methods working on bigrams and trigrams Also, maybe we are supposed to permute the number string before decoding... Take every 15th number or something... Link to comment Share on other sites More sharing options...
Aminar Posted March 5, 2014 Report Share Posted March 5, 2014 Well, using 10 as a code didn't end up with any proper splits as all. And transposition -or simply mixing the letters invalidate statistical methods working on bigrams and trigrams Also, maybe we are supposed to permute the number string before decoding... Take every 15th number or something... It isn't long enough for much of that... Link to comment Share on other sites More sharing options...
Satsuoni he/him Posted March 5, 2014 Author Report Share Posted March 5, 2014 I mean, take every 15th number then rotate once you reach the end of the string (treat it as ring). At least some numbers should span all the digits before repeating when taken this way. Unlikely, though, given the structure present in string. Link to comment Share on other sites More sharing options...
Voidbringer Posted March 5, 2014 Report Share Posted March 5, 2014 The numbers 10, 11, 12 appear in succession several times. 1 Link to comment Share on other sites More sharing options...
Aminar Posted March 5, 2014 Report Share Posted March 5, 2014 (edited) The numbers 10, 11, 12 appear in succession several times. Nice Catch. That puts 10 11 12 as probably a basic word. Moreover I'm almost sure 11 is a vowel. Two of those 10 11 12's are followed by 3 4. There's also a 5 5 11 5 3 4 appears 4 times. This implies it's a common ending to words.... 5 5 12 12 Both of these possibly appear. Double letter combos are rareish. rr tt pp ss dd ll bb nn mm oo ee Edited March 5, 2014 by Aminar 1 Link to comment Share on other sites More sharing options...
Satsuoni he/him Posted March 5, 2014 Author Report Share Posted March 5, 2014 (edited) For what it is worth, 1234 also appears two times, before and after 5511, but that might be just a coincidence Edit: and 149149 make it likely that 9 is a single code digit Edited March 5, 2014 by Satsuoni 1 Link to comment Share on other sites More sharing options...
Aminar Posted March 5, 2014 Report Share Posted March 5, 2014 Just borrowed this from the internet. 12 => 1510 => 1311 => 1151 => 1001 => 915 => 841 => 721 => 734 => 625 => 671 => 591 => 514 => 5 6 letter sequences101112 => 4 3 4 Showing up six time could mean a lot. I cut the 3 4 and 5 letter sequences because they aren't likely to matter. I wish there was a good way to tell when something is 11 or 17 or 1 7 and the like. Link to comment Share on other sites More sharing options...
Bibliovortex Posted March 5, 2014 Report Share Posted March 5, 2014 I don't think this code is based on a 26-letter alphabet substitution, and here's why - the distribution is all wrong. There are 8 numbers that would have a 2 in them, and 11 that would have a 1; even accounting for letter frequency, there still shouldn't be such a massive difference between the number of 1's and the number of 2's (almost fourfold). My first thought was that it might be based on 10, since that number is so important to Roshar, but that leaves far too many 1's even if you exclude the ones that are really 10's. When I looked at Aminar's distribution chart again, though, I noticed that the digits 0-6 occur much more frequently than the digits 7-9. If you leave out the 1's, 0 and 2-6 occur an average of 11 times each. 7-9 occur an average of 4.33 times each. This, combined with the very high number of 1's, implies to me that our substitution has 16 letters instead of 10. It explains why there are so few 2's compared to 1's (and so many 1's in general), and why there are so few 7's, 8's and 9's. I think it's at least plausible to go on the assumption that Taravangian had, on this day, some level of cosmere awareness. (See the epigraph to chapter 85 - sounds to me like he's talking about a Worldhopper.) So it isn't unthinkable that he might come up with a code based on the number 16. The only question is, what 16 letters are being used in the substitution? I don't have the math or programming chops to try to get a brute force answer to that question, so I'll leave it for somebody else to try that. If the 10-11-12 theory is right, the word in question is quite possibly "the," which is something to start with at least... 3 Link to comment Share on other sites More sharing options...
Aminar Posted March 5, 2014 Report Share Posted March 5, 2014 (edited) I don't think this code is based on a 26-letter alphabet substitution, and here's why - the distribution is all wrong. There are 8 numbers that would have a 2 in them, and 11 that would have a 1; even accounting for letter frequency, there still shouldn't be such a massive difference between the number of 1's and the number of 2's (almost fourfold). My first thought was that it might be based on 10, since that number is so important to Roshar, but that leaves far too many 1's even if you exclude the ones that are really 10's. When I looked at Aminar's distribution chart again, though, I noticed that the digits 0-6 occur much more frequently than the digits 7-9. If you leave out the 1's, 0 and 2-6 occur an average of 11 times each. 7-9 occur an average of 4.33 times each. This, combined with the very high number of 1's, implies to me that our substitution has 16 letters instead of 10. It explains why there are so few 2's compared to 1's (and so many 1's in general), and why there are so few 7's, 8's and 9's. I think it's at least plausible to go on the assumption that Taravangian had, on this day, some level of cosmere awareness. (See the epigraph to chapter 85 - sounds to me like he's talking about a Worldhopper.) So it isn't unthinkable that he might come up with a code based on the number 16. The only question is, what 16 letters are being used in the substitution? I don't have the math or programming chops to try to get a brute force answer to that question, so I'll leave it for somebody else to try that. If the 10-11-12 theory is right, the word in question is quite possibly "the," which is something to start with at least... Issue being that this is a small sample size no matter what we work off of. The numbers aren't likely to be nice and even based on such a small sample. And there are more than 16 necessary pairs throughout the sample... Working off of the 10 11 12 being The H 18 25 THE 7 E 4 9 15 E T T H 14 T 21 5 1 17 1 H THE 17 1 3 4 4 8 3 H T 7 15 14 25 4 14 3 4 T 9 16 14 9 14 9 3 4 12 12 25 4 T T 12 5 12 7 T 15 19 T H E 3 4 E 5 5 H 5 25 E 15 7 5 5 H 12 3 4 T H E 9 15 E T 6 15 3 4 This has me thinking 14 might be A H 18 25 THE 7 E 4 9 15 E T T H A T 21 5 1 17 1 H THE 17 1 3 4 4 8 3 H T 7 15 A 25 4 A 3 4 T 9 16 A 9 14 9 3 4 12 12 25 4 T T 12 5 12 7 T 15 19 T H E 3 4 E 5 5 H 5 25 E 15 7 5 5 H 12 3 4 T H E 9 15 E T 6 15 3 4 And 15 might be B or G? Also n 18 25 a n d 7 d 4 9 15 d a a n 14 a 21 5 1 17 1 d a n d 17 1 3 4 4 8 3 n a 7 15 14 25 4 14 3 4 a 9 16 14 9 14 9 3 4 d d 25 4 a a d 5 d 7 a 15 19 a n d 3 4 12 5 5 n 5 25 d 15 7 5 5 n d 3 4 a n d 9 15 d a 6 15 3 4 Edited March 5, 2014 by Aminar Link to comment Share on other sites More sharing options...
Recommended Posts