Beale Cryptanalysis

[Pistolshrimp]

Hello all. Not sure if this is the right place for this. It's kind of a long post so I apologize in advance. Most of the length comes from the Methods section trying to explain what I am doing, if you wish you can skip that and still get the jist of the post. I tried to put it into a Spoiler tag but I don't see that option here so I put it in a quote so it can be skipped easily if so desired.

I am a GIS and Programming student who often finds myself drawn to mysteries and trying unique ways of solving them. While interesting I use them as ways to build skills, force me to think outside of the box, and ultimately to build my resume.

I am not a historian or a treasure seeker. I am not necessarily trying to crack the ciphers. That is probably outside my ability, and judging from the posts here, someone seems to "crack" them every couple of months. No shade of course, just that it is probably not worth the effort for me to go too far down that rabbit hole. I am posting on this site because it seems like you all are really knowledgeable and, perhaps for internet forums, uncharacteristically supportive. i have some conclusions at the bottom I would like to run by you all and see if you disagree or if there is something out there proving me wrong.

METHODS/EXAMPLES

What I have been doing is some cryptanalysis on the codes themselves. I built a program which inserts sentences into each of the ciphers and counts every time it fits and subsequent Trigrams it then creates. The sentences are made from an external file so I can essentially put in any text and it creates every sentence possible of lengths 1 - 10 for words adjacent to each other.

For instance take, "I HAVE DEPOSITED IN THE COUNTY OF BEDFORD ABOUT FOUR". My program would create a sentence for each word, then make sentences for "I HAVE", "HAVE DEPOSITED", "DEPOSITED IN" and so on until it would use the whole thing. It would do this for all the words in the entire text.

Then it would try to fit every sentence into the cipher and count the times it fits and the amount of accidental Trigrams it creates. Trigrams being three letters in a row that are common in other words and which may or may not be other words themselves, for instance ("THE", "AND", "NTH", "ERE").

Lets take my earlier sentence as an example: The entire word block, "I HAVE DEPOSITED IN THE COUNTY OF BEDFORD ABOUT FOUR" fits into the Third Cipher 25 times, changing every other number to the assigned letter. It only won't fit when it would have to re-write over a letter, IE when the program tries to assign two different letters to the same number. Of those 25 successful fits only 3 of them create extra common Trigrams.

For here is one return: Trigams found = 'TIO' and 'ORT'.
Puesdo PlainText:
IHAVEDEPOSITEDINTHECOUNTYOFBEDFORDABOUTFOURS****O***D**TIO****O*****E*********O**H**DI**T*****D**DS***OD****SAO**O*TC**HUF*NB**OHO*FD***E********************FT*E****U***S*F********************E********FE****S*O*P**U**D***VTB********D***IH*****A***O**I******UDH*****O***TRA********FDA*O**S*A***UE**********Y**DH*T**B****H*****D**************DODF*****S**B*****F************OED**E*SBORT*******H************F*O*D****BVDS**F****************E*B***D***********N*****VA*B****DS************S**BT*F**Y**U*E**O********F**********A*******NE************************DO****BD*EO*******ED****D*F**S**********************U*************

As you can see while this does make Trigrams and things that look like they could be English words it also produces nonsense strings like "BVDS". My program does not filter out these nonsense stings. I typically don't even look at the plaintext, I just printed it out to better explain the process.

What is interesting, at least to me, is if I take the same sentence and run it though Cipher 2. This is essentially my control.

Attempts that include 'I HAVE DEPOSITED IN THE COUNTY OF BEDFORD ABOUT FOUR':2
Attempts which include a Trigram not already in Key: 1
Trigrams: ['AND', 'ION', 'VER', 'ONE', 'INE', 'UND']
Puesdo PlainText:
IHAVEDEPOSITEDINTHECOUNTYOFBEDFORDABOUTFOUR***E*FRO***FO***I***E***VA***N*RV*U*****F****E****HE**RF***O******O******F***O**N***TI*****E*ON*I***O****YT***E*****E**********S**E*IV********E***R**H**E*******F****D**O***C***I*TC***T*NHU*DRE****F*U**EE**O**D******D*****I*TY*I*H**U***ED**D**E*VE**UND***SI*VER***OS**E*NOV******E**INE*EEN**E****N******D**E******E**T*EN*YO****D*ON*I*************N****RED*N***VE**O*ND*OF*O*****T***V**U***ED******H*YE**H******V*R*******E*******N**I**T*O****NE******ETO**V************ION***V********HI*TE**R****AND*O*******EA**VE**S**U***Y**C**D*N**O**O***********OV****H*V*******OU***Y***ED******ONEA*D*H*VE*****RE*TON*O*ID**O*E******C*V********O*H******ER****ER*NE*****I****HC******O*****Y*F***V*****O***T*****FIC***Y***********F*N*IN*I*

Only 1 solution works that provides readable common Trigrams - the real solution.

Again I am not looking at the plaintext at all, I am running thousands of sentences through each of the ciphers and looking for trends.

INITIAL DATA:
Depending on the length of the input text the program may takes several hours to run so I only have a couple of runs.

I ran 2 different texts, the first the successful solution to cipher 2, and for the second I tried to find some text that seemed as far as I could get from what is contained in the ciphers: the first two paragraphs of a review of one of Ariane Grande's albums.



I have highlighted the part that I think is most interesting. The Trigrams found.



Here's are a couple of graphs that may help visualize things. (I apologize for the very poor quality graphs, I should have normalized the data and worked on colors to increase readability.)

So interestingly cipher 2 creates a high common Trigram count even when using the wrong key to decipher it. (Unless Beale was an Ariana Grande fan and hid a secret review of her album in his cipher.) Additionally they both follow the same slightly left leaning bell curve. Cipher's 1 and 3 also produce similar curves no matter what text is put thought (obviously only 2 data points at the moment).

Conclusions I am leaning towards.
1. Ciphers 1 and 3 may not be real. They do not follow the expected pattern set by cipher 2.
2. If ciphers 1 and 3 are real they differ quite a bit from cipher 2 in construction. With 1 differing much more than 1.
3. If cipher 1 is real this high difference is MAY be accounted for by not being English sentences but rather lists of names. Implying cipher 1 is actually cipher 3.

I am writing all this because you all are probably the experts on Beale and hope to get your thoughts. None of this is conclusive; I need to do more runs and compute a p-value to see if these results are greater than chance, but so far they look like they are.

What are your thoughts on my conclusions?

 

Amazon Forum Fav 👍

DR.ÖTEK Metal Detector, Waterproof Gold Detector, Pinpoint & Terrain Modes, Professional Metal Detector - Grab it through Amazon!

[Rebel - KGC]

Hello all. Not sure if this is the right place for this. It's kind of a long post so I apologize in advance. Most of the length comes from the Methods section trying to explain what I am doing, if you wish you can skip that and still get the jist of the post. I tried to put it into a Spoiler tag but I don't see that option here so I put it in a quote so it can be skipped easily if so desired.

I am a GIS and Programming student who often finds myself drawn to mysteries and trying unique ways of solving them. While interesting I use them as ways to build skills, force me to think outside of the box, and ultimately to build my resume.

I am not a historian or a treasure seeker. I am not necessarily trying to crack the ciphers. That is probably outside my ability, and judging from the posts here, someone seems to "crack" them every couple of months. No shade of course, just that it is probably not worth the effort for me to go too far down that rabbit hole. I am posting on this site because it seems like you all are really knowledgeable and, perhaps for internet forums, uncharacteristically supportive. i have some conclusions at the bottom I would like to run by you all and see if you disagree or if there is something out there proving me wrong.

METHODS/EXAMPLES

What I have been doing is some cryptanalysis on the codes themselves. I built a program which inserts sentences into each of the ciphers and counts every time it fits and subsequent Trigrams it then creates. The sentences are made from an external file so I can essentially put in any text and it creates every sentence possible of lengths 1 - 10 for words adjacent to each other.




Again I am not looking at the plaintext at all, I am running thousands of sentences through each of the ciphers and looking for trends.

INITIAL DATA:
Depending on the length of the input text the program may takes several hours to run so I only have a couple of runs.

I ran 2 different texts, the first the successful solution to cipher 2, and for the second I tried to find some text that seemed as far as I could get from what is contained in the ciphers: the first two paragraphs of a review of one of Ariane Grande's albums.

View attachment 1832110

I have highlighted the part that I think is most interesting. The Trigrams found.

View attachment 1832102

Here's are a couple of graphs that may help visualize things. (I apologize for the very poor quality graphs, I should have normalized the data and worked on colors to increase readability.)

So interestingly cipher 2 creates a high common Trigram count even when using the wrong key to decipher it. (Unless Beale was an Ariana Grande fan and hid a secret review of her album in his cipher.) Additionally they both follow the same slightly left leaning bell curve. Cipher's 1 and 3 also produce similar curves no matter what text is put thought (obviously only 2 data points at the moment).

Conclusions I am leaning towards.
1. Ciphers 1 and 3 may not be real. They do not follow the expected pattern set by cipher 2.
2. If ciphers 1 and 3 are real they differ quite a bit from cipher 2 in construction. With 1 differing much more than 1.
3. If cipher 1 is real this high difference is MAY be accounted for by not being English sentences but rather lists of names. Implying cipher 1 is actually cipher 3.

I am writing all this because you all are probably the experts on Beale and hope to get your thoughts. None of this is conclusive; I need to do more runs and compute a p-value to see if these results are greater than chance, but so far they look like they are.

What are your thoughts on my conclusions?

WOW!

 

[franklin]

Hello all. Not sure if this is the right place for this. It's kind of a long post so I apologize in advance. Most of the length comes from the Methods section trying to explain what I am doing, if you wish you can skip that and still get the jist of the post. I tried to put it into a Spoiler tag but I don't see that option here so I put it in a quote so it can be skipped easily if so desired.

I am a GIS and Programming student who often finds myself drawn to mysteries and trying unique ways of solving them. While interesting I use them as ways to build skills, force me to think outside of the box, and ultimately to build my resume.

I am not a historian or a treasure seeker. I am not necessarily trying to crack the ciphers. That is probably outside my ability, and judging from the posts here, someone seems to "crack" them every couple of months. No shade of course, just that it is probably not worth the effort for me to go too far down that rabbit hole. I am posting on this site because it seems like you all are really knowledgeable and, perhaps for internet forums, uncharacteristically supportive. i have some conclusions at the bottom I would like to run by you all and see if you disagree or if there is something out there proving me wrong.

METHODS/EXAMPLES

What I have been doing is some cryptanalysis on the codes themselves. I built a program which inserts sentences into each of the ciphers and counts every time it fits and subsequent Trigrams it then creates. The sentences are made from an external file so I can essentially put in any text and it creates every sentence possible of lengths 1 - 10 for words adjacent to each other.




Again I am not looking at the plaintext at all, I am running thousands of sentences through each of the ciphers and looking for trends.

INITIAL DATA:
Depending on the length of the input text the program may takes several hours to run so I only have a couple of runs.

I ran 2 different texts, the first the successful solution to cipher 2, and for the second I tried to find some text that seemed as far as I could get from what is contained in the ciphers: the first two paragraphs of a review of one of Ariane Grande's albums.

View attachment 1832110

I have highlighted the part that I think is most interesting. The Trigrams found.

View attachment 1832102

Here's are a couple of graphs that may help visualize things. (I apologize for the very poor quality graphs, I should have normalized the data and worked on colors to increase readability.)

So interestingly cipher 2 creates a high common Trigram count even when using the wrong key to decipher it. (Unless Beale was an Ariana Grande fan and hid a secret review of her album in his cipher.) Additionally they both follow the same slightly left leaning bell curve. Cipher's 1 and 3 also produce similar curves no matter what text is put thought (obviously only 2 data points at the moment).

Conclusions I am leaning towards.
1. Ciphers 1 and 3 may not be real. They do not follow the expected pattern set by cipher 2.
2. If ciphers 1 and 3 are real they differ quite a bit from cipher 2 in construction. With 1 differing much more than 1.
3. If cipher 1 is real this high difference is MAY be accounted for by not being English sentences but rather lists of names. Implying cipher 1 is actually cipher 3.

I am writing all this because you all are probably the experts on Beale and hope to get your thoughts. None of this is conclusive; I need to do more runs and compute a p-value to see if these results are greater than chance, but so far they look like they are.

What are your thoughts on my conclusions?

Start your sentence out with, "The treasure is due north of the inn. I left Paschal's going north for about four miles and then out from Read's Point......." I gurantee you will not only find words but sentences using your method. Some of it will be off because once in a while two ciphers should be one cipher and one cipher can be two ciphers. But you should get at least 60 percent of the letters to fall into place. Let us see what you come up with?

 

[Mudflap]

Interesting.

 

[jhonnz41]

Interesting.

Cipher 1 is real, I am sure computer cant solve this unless you can use a bit complex algorithm.. Cipher 3 is still gibberish to me, I searched for same numbers close with each other to try to substitute them letters that I think would fit based on the names that I have..But perhaps they have different key, or maybe these letters could be found in the first or last word 10 words prior to the number..

 

[Pistolshrimp]

Start your sentence out with, "The treasure is due north of the inn. I left Paschal's going north for about four miles and then out from Read's Point......." I gurantee you will not only find words but sentences using your method. Some of it will be off because once in a while two ciphers should be one cipher and one cipher can be two ciphers. But you should get at least 60 percent of the letters to fall into place. Let us see what you come up with?

I appreciate your response. Testing your starting phrase did help me find a couple of errors in my code that I was able to fix. (Primarily it wasn't building as many sentences as it should have which will deflate my overall results.)

With that said putting "THE TREASURE IS DUE NORTH OF THE INN. I LEFT PASCHALS GOING NORTH FOR ABOUT FOUR MILES AND THEN OUT FROM READS POINT" yields:

THETREASUREISDUENORTHOFTHEINNILEFTHASCHALSGOINGNORTHFOAABHUTFOURMIEESANDTHENOUFFROMRCALSNOINT**S*LT***DF*O*****I**I*T**U***AO*****E*EAO*****MTT*************U****O*****IU*****A*O*LD*O****A*L*T*****I*B*****************ESRT*****E**T***A******T*R***FO***O*******U***O**R***N*****O******T***T*T****H**O****DA*O******O***E*********T*U*********O*****T*********F*****OD***E****A****S****RO******T********E******N*****O*E***H*****A*S****I**A*D*AMH*FO**N**ASA*R*N****SH***R*ND**DT*********************OE********FET****O**TN****A**

First off you will notice that it doesn't fit. You mention that is because sometimes a single number will have 2 letters. My program, at least how it is written, cannot handle this.

I say this with all due respect as an outsider to the Beale papers and this forum, but it seems complicated to assume that a single letter will sometimes have 2 letters. This, of course, doesn't mean it isn't the case. Beale could very well have definitely required it. As a programmer the way I would approach it is that if I thought there was a need to have a single number be more than one letter I would try to find a pattern to it. From my understanding for people to take a solution seriously at this point in time, since so many people have claimed to break it, the key must be able to be conveyed to Mr. Morris in a way he could recreate it. There would have to be some pattern that Mr. Morris could use. For instance the first repeating Number is "40" which is assigned to "H" which makes "I LEFT HASCHALS GOING NORTH". Beale could have wrote "Add 8 to every other repeating number". This obviously doesn't hold up, but if it was something like that I could program it and look for it. I think as an outsider I would be more likely to believe a deciphered code the less abstraction it required to decipher. IE 0 Errors or double keys > 1 error or double keys > an easy double key system > a complicated double key system ... etc.

Your start does produce a lot of interesting word pieces which could very well mean that it is the start of a correct translation. However it doesn't produce any common Trigrams.

My Trigram List = ["THE", "AND", "ING", "HER", "THA", "ENT", "ERE", "ION", "ETH", "NTH", "HAT", "INT", "FOR", "ALL", "STH", "TER", "EST", "TIO", "HIS", "OFT", "HES", "ITH", "ERS", "ATI", "OTH", "FTH", "DTH", "VER", "TTH", "THI", "REA", "SAN", "WIT", "ATE", "ARE", "EAR", "RES", "ONT", "TIN", "ESS", "RTH", "WAS", "SOF", "EAN", "YOU", "SIN", "STO", "IST", "EDT", "EOF", "EVE", "ONE", "AST", "ONS", "DIN", "OME", "CON", "ERA", "STA", "OUR", "NCE", "TED", "GHT", "HEM", "MAN", "HEN", "NOT", "ORE", "OUT", "ORT", "ESA", "ERT", "SHE", "ANT", "NGT", "EDI", "ERI", "EIN", "NDT", "NTO", "ATT", "ECO", "AVE", "MEN", "HIN", "HEA", "IVE", "EDA", "INE", "RAN", "HEC", "TAN", "RIN", "ILL", "NDE", "THO", "HAN", "COM", "IGH", "AIN", "UND", "OUN", "DAN", "LVE", "NDS", "DTO", " WIL", "PAR", "ART", "HEP"]

This list is produce by taking the 100 most common English trigrams, the 50 most common ones in the entirety of the Beale Papers and the top 10 from the deciphered paper 2. As there is considerable overlap there is less than the expected 160.

I know that every deciphered code was a work of passion so I don't want to come off as dismissive or discouraging. Again my program doesn't really disprove anything but rather looks at trends. I have no problem with running phrases tho. I have done it a bunch myself for fun.

My favorite result is uses: "SOUTH DOWN GOOSE CREEK"

*****************************H*********C********************W*********D****U********SC*****E****H***O***W***O*************S*********S****************************************SOUTHDOWNGOOSECREEK*********O************************R****S*****************W***************K ********************************ONDS*U*************************************O****************W*****K*******************************************************S*****N************W*****************************K*****************************O*******************

It doesn't yield a hole lot but it does have an interesting quadragram "ONDS" and no letter combos that would imply a bad fit.

In any case I need to re-run my data now that I fixed a couple of errors.

 

[Pistolshrimp]

Cipher 1 is real, I am sure computer cant solve this unless you can use a bit complex algorithm.. Cipher 3 is still gibberish to me, I searched for same numbers close with each other to try to substitute them letters that I think would fit based on the names that I have..But perhaps they have different key, or maybe these letters could be found in the first or last word 10 words prior to the number..

Again, I am more cautious of a solution the more complex the key would have to be for it to be deciphered. As I understand you are currently trying to secure your legal rights perhaps in one of your threads you can explain your methodology without necessarily revealing the actual plaintext.

 

[jhonnz41]

Again, I am more cautious of a solution the more complex the key would have to be for it to be deciphered. As I understand you are currently trying to secure your legal rights perhaps in one of your threads you can explain your methodology without necessarily revealing the actual plaintext.

In the page 1, the contents of the message has parts...The He gave Location, Description, instruction, Landmarks, then back to Location, description and Instruction.. The page 1 is far more complex than the page 2, I don't know why they did that...perhaps they really want to secure the message..A cipher within a cipher, but if you can figure it out, you simply just write it, no sweat at all and it will form words, then into sentences, then into paragraphs..

 

[Rebel - KGC]

If you say so...

 

[jhonnz41]

If you say so...

They went north at first to look for a cave..Saltpetre Cave, but it was unfit for them..So the went further north..

 

[franklin]

They went north at first to look for a cave..Saltpetre Cave, but it was unfit for them..So the went further north..

Saltpetre Cave would have been the last place they would consider to bury a treasure. You must remember saltpetre was in high demand by not only the locals but for the military. They used it to make gunpowder.

 

[jhonnz41]

Saltpetre Cave would have been the last place they would consider to bury a treasure. You must remember saltpetre was in high demand by not only the locals but for the military. They used it to make gunpowder.

I did not mention they buried it there..:-D why would I mention the exact spot...

They went further north of Saltpetre cave, they saw Gold mines but it was not in that spot...The treasure is still in bedford..
I am revealing this to tell you his route...
You guys always focused on the prize, but I focused on some of the details based on descriptions...

 

[Rebel - KGC]

HA! NO Gold Mines NORTH of there; only IRON ORE MINES. Area was once known as IRONVILLE now known as Villamont, Va.

 

[ECS]

Seems someone needs to iron out the kinks in their decipherment.
Maybe another Jean Laf consulting sessions is required to smooth out the message, he is, after all by his own admission the ONLY one to correctly solve the Beale ciphers, and its been said a "real nice guy", or Eldo or Legrand, two other real nice guys who also claimed to be the ONLY ones to correctly solve the Beale codes.
Speaking of Legrand, whatever happened with Icehouse's BEALE MYSTERY TEAM and the big dig of the "found" Beale vault"

 

[Rebel - KGC]

Seems someone needs to iron out the kinks in their decipherment.
Maybe another Jean Laf consulting sessions is required to smooth out the message, he is, after all by his own admission the ONLY one to correctly solve the Beale ciphers, and its been said a "real nice guy", or Eldo or Legrand, two other real nice guys who also claimed to be the ONLY ones to correctly solve the Beale codes.

 

[jhonnz41]

I need help...I need more information. Virginia Museum wants me to physically go in their museum to get this data....Someone point me Gold mines near Fincastle...

 

[jhonnz41]

HA! NO Gold Mines NORTH of there; only IRON ORE MINES. Area was once known as IRONVILLE now known as Villamont, Va.

Maybe Beale assumed that it was a Gold mine that they saw?

 

[jhonnz41]

Weird, Base on the number of miles that he gave it feels right..2 of the distances Beale gave which was confirmed by the museum is accurate...So I just assumed this data....I wish I know more since I am more interested of his route..

 

[jhonnz41]

Also, the Saltpetre cave that they want to originally hide their treasure was exactly the same number of miles given in his message from a specific location...Also same descriptions...

 

[jhonnz41]

Wait, No...I did not mean Villamont Va...what I mean is maybe somewhere near Rich patch mines and Jordan mines...cause he said Gold mine(s) which means more than one...Or he just assumed it was Gold mines?