But debunking IMO only gets us so far. It can maybe show a wrong path, but can it show a right path? Is there a point in having a cipher section for new ideas? Don’t at some point you have to TRY something? Take a risk of being wrong instead of just showing someone else is wrong?

"Trying new ideas" and "debunking" are not mutually exclusive. Debunking only happens when ideas fail to stand up when tested.

If I have cancer, I want scientists to try new ideas and new drugs. But I only want to take the drugs that pass the tests!

The purpose of a scientific test is not to discourage new ideas: It’s to eventually sharpen your focus onto the ideas that matter. So keep trying new ideas! One might pass the tests.

AK-

Sometimes I wonder if you are paying attention, or even read what others have written. Quoting you from just above:

Pretend there is no Graysmith.
Start with a blank slate. Using parts of the Harden Key, ETA frequency analysis and some word solve "guesses".
Starting with a blank slate, I think QT”s study, backed up by my analysis, and even work from doranchak and glurk, shows that the most likely candidate (or at least a very good candidate) for + is L.
So lets TRY the + as L. Add to that S99 as SEE. Solve for TH. If you proceed from there with further frequency analysis and possible word solve guesspothesis, you will usually end up with something looking at least in part like the Raw Graysmith.

Don’t you see the image that doranchak just posted? The one that HAS the + symbol as "L" – The one that HAS the S99 as "SEE"
You asked me to lock those letters into the ZKD program, and see what it would come up with. I did. That is what that image is.
The same one that doranchak showed contained many words forwards, backward, diagonally, and even "THEO"

I DID try what you asked. You were just looking at it.

-glurk

Yes I know you tried it once. We discussed it. And I appreciate it.

I think it should be tried more than once. Tried with SEE A NAME, tried without it. Tried with other ideas, your ideas, other possible word solves or symbol solves.

What you tried was only solved for SEE, the + and for a TH combo. Thats it.

I am suggesting trying with more possible solves added, like SEE A NAME, and the 3rd line THEO and 9 letters around it. You may or may not think it worth trying, others might, and I will.

I do appreciate your effort and your ideas. I am offering this work as a starting point to anyone looking for one. Some have expressed interest.

Doranchak: "Trying new ideas" and "debunking" are not mutually exclusive. Debunking only happens when ideas fail to stand up when tested.

If I have cancer, I want scientists to try new ideas and new drugs. But I only want to take the drugs that pass the tests!

The purpose of a scientific test is not to discourage new ideas: It’s to eventually sharpen your focus onto the ideas that matter. So keep trying new ideas! One might pass the tests.

AK Wilks: That I agree with you on.

This is what I think is worth TRYING as a starting point, to see if it gets us anywhere. It may or may not.

If the code is "Back slang" cool it would be:
http://en.wikipedia.org/wiki/Backslang
BACK SLA ng ….

AK,

Based on an idea from glurk, I ran an experiment that tries to find substitution keys that result in many "Zodiac-like" words appearing in the plain text in various directions. Here is a small sampling of the results:

http://zodiackillerciphers.com/wordsearch-examples/zodiac-words.html

My conclusion from this is that there are many valid decodings that contain interesting words appearing in different directions in the plain text. So, I think it’s impossible to pick one of them and declare it as being the right one, unless the rest of the plain text is readable.

Thanks. Interesting. No doubt we can intentionally create a word puzzle solution with vertical and diagonal words. This proves little to me. Maybe I don’t understand what you were trying to show?

Why not take six or so actual proposed solutions, Starliper and others you looked at, plus my Raw GraySmith, and search for correct spelling nonanagram words of four or more letters appearing vertical, diagonal and backwards. Lets see if there is a significant difference.

Thanks. Interesting. No doubt we can intentionally create a word puzzle solution with vertical and diagonal words. This proves little to me. Maybe I don’t understand what you were trying to show?

I’m trying to demonstrate how easy it is for words to appear by chance. The reason it’s easy to intentionally create word puzzle solutions is that many words will appear in any sample of text written out in a 17×20 block.

Why not take six or so actual proposed solutions, Starliper and others you looked at, plus my Raw GraySmith, and search for correct spelling nonanagram words of four or more letters appearing vertical, diagonal and backwards. Lets see if there is a significant difference.

That’s a good suggestion. Here are the results:

• The Graysmith solution
• Starliper’s solution
• The known solution for the 408
• John Cecil’s solution
• Hal Kravcik’s solution
• Daryll Lathers’ solution
• Traveller1st’s solution
• Ozcan Turkmen’s solution
• Your own text from the above post
• Completely random text

[/list:u:202un27y]

Each link takes you to the words found in each grid of plain text. Words are highlighted in the grids for easy viewing.

The upshot is that it’s an uphill battle to prove that the extra words found in the Graysmith solution were placed there by design. This experiment seems to show that they are appearing only by chance.

http://zodiackillerciphers.com

 

Thanks!

Well its Oscar time so just a quick look for now but this will be examined by me in detail tomorrow. Trav, up2something, glurk, others, what say you?

But I think all the proposed solutions have 40-56 word counts most between 40 to 45. They are 42, 56, 40, 46, 42 and 49. I think the average is 45.8.

The Raw Graysmith has 94.

That’s more than TWICE as much as the average proposed solutions! 200% of the average. That’s a significant deviation. Isn’t it? And it has more Zodiac relevant words than any of them. However you may have over included on Graysmith I count 23 words that are horizontal left to right. Why were those included? Without them the Raw Graysmith has 71 words. Even that is more than 150% of the average which is statistically significant.

Sorry for the confusion – the Graysmith one included left-to-right words, and the others didn’t. Check the links again – I’ve updated them all to include left-to-right words.

I think your premise is that:

1) The plain text is a legal substitution of the cipher symbols (i.e., no polyphones allowed)
2) A "good" solution has many non-anagrammed Zodiac-related words appearing backwards, diagonally, and vertically.

The question I’m now interested in is: How many valid substitutions of the 340 produce plaintexts that contain at least as many Zodiac-related words as the Graysmith solution? I will run an experiment to try to get a feel for this.

OK that may be an interesting experiment.

But my original query was does the Raw Graysmith proposed solution have signicantly more words of four letters or more appearing vertically, diagonally and backwards than other proposed solutions?

The answer is a resounding YES by a margin of 55% more than the average we would expect. Secondarily, it also has way more Zodiac and/or crime related words.

I think this a black swan event that should not be rationalized or ignored but rather is worthy of further study. Thanks again for doing this. I am not quite sure what it means yet but I do think the deviation is significant and worthy of a deeper look.

Thanks again to Doranchak for doing this very interesting study.

Based on my initial look at this data I think it tends to confirm that the Raw Graysmith has significantly more words appearing correct spelling non-anagram vertically, diagonally and backwards then would happen by mere chance. The six other proposed solutions studied by Doranchak have an average of 45.8 words appearing, and many are nonsense words or rarely used, while the Raw Graysmith has 69. That is more than 50% larger than the average, which is a significant deviation not likely to happen by chance. It is an outlier for sure. What does it mean? It also has 11 words that would be considered relevant to the Zodiac case or crime in general, while the average for the others is about 3.66.

Here is my revised count of the Raw Graysmith as shown by Doranchak’s computer search program:

1 six letter word
7 five letter words
61 four letter words

Total of 69 words appearing vertically, diagonally or backwards
At least 11 are words relevant to the Zodiac case or crime in general (BOMBS, SLASH, TELL, LIES, DUEL, GAME, LIST, BOMB, TIES, BARS, LASS)

Here is my revised count of one typical proposed solution, the Starliper, as shown by Doranchak’s computer search program:

2 five letter words
37 four letter words

Total of 39 words appearing vertically, diagonally or backwards
At least 4 are words relevant to the Zodiac case or crime in general (TIME, HELP, EVIL, LIED)

All the proposed solutions total words, Zodiac/Crime relevant:

Cecil: 60, 4 (FAKE, FAKERS, MAIL, FATE)
Starliper: 39, 4 (TIME, HELP, EVIL, LIED)
Kracik: 39, 3 (DEAL, SINS, LIED)
Trav1st: 42, 5 (SEEN, DATE, SHOT, SECT, HINT)
Lathers: 46, 4 (LANE, TIED, SINS, LASS)
Turkmen: 49, 2 (LIES, FOIL)
Raw Graysmith: 69, 11 (BOMBS, SLASH, TELL, LIES, DUEL, GAME, LIST, BOMB, TIES, BARS, LASS)

The Raw Graysmith has a 50% higher total word count than the average, and a 300% higher Zodiac/Crime Relevant word count.

That is more than 50% larger than the average, which is a significant deviation not likely to happen by chance. It is an outlier for sure. What does it mean? It also has 11 words that would be considered relevant to the Zodiac case or crime in general, while the average for the others is about 3.66.

I think you have made an overly hasty conclusion based on a very limited data sample. It is easy to find solutions that:

1) Are valid substitutions of the cipher text
2) Contain a dozen or so Zodiac or crime related words
3) Contain numerous other words

Here are four such solutions:

• Solution 1 – BATES GIRL NEWS FIRE LIST HEAT LIPS HELL LOSE WILD HOLE LIES SHOT BOMB BARS ROPE SHIT HANG TIES THEO TEEN CAEN
• Solution 2 – STALLS ALLEY TAKE GIRL PAUL BALL HEAT HELL HOLE LIES SICK LAKE SHIT RAGE RAGE TIES BANG STAB DUEL TEEN
• Solution 3 – PINES TAKE ROAD GAME LIST PAUL PAIN SKIN PARK HEAT HELL LOSE LIES KILL SHOT BARS GUNS HANG FOOL RAGE TIES PIGS DUEL TEEN
• Solution 4 – DEATH BLOND TAKE GAME DEAD BALL HEAT HELL LOSE LIES LAKE BARS ROPE SHIT RAGE TIES BANG HERB STAB CAEN

[/list:u:38kehnl7]

Each solution has an average of 136 words, and an average of 21 Zodiac or crime related words. Both averages are well above those for the Graysmith solution. Therefore, the Graysmith solution can’t be said to be an outlier.

http://zodiackillerciphers.com

 

"Overly hasty conclusions". Wait I have not reached any conclusions yet! I have prefaced all my posts by saying this is an initial look that needs to be studied further.

Where did these solutions come from? You created them? With what parameters and intent? The intent to create vertical and diagonal words?

The study that I asked you to do, and which you did very well, was to look at six other proposed solutions and compare them to the Raw Graysmith (as deduced by Ed O’Neil/Kite/Wilks) and see if the RG had less, the same or more 4 letter or more words appearing vertically, diagonally and backwards.

By my initial study it does. By a margin of 50% more.

I am interested for your take on that. Isn’t 50% usually considered a significant deviation? What is your preliminary explanation? Do you think it merits further study?

And while my Zodiac/Crime words designation is by nature somewhat subjective, I tried to fair. Why is the average for the 6 other proposed solutions 3.66 Z/C words while for the RG it is 11? That is 300% more.

Why do I think this MAY mean something? Of course we can intentionally create a solution with many Z/C words and other words appearing in a word puzzle way. What I claim may be significant is that Graysmith did NOT do this. To my knowledge he never published a raw solution, and in all his books and writings he never claimed "Hey BOMBS and LIST appear diagonally", and to this day has never noticed it or commented on it. The same holds true for Ed O’Neil, who deduced what the actual raw solution was before Graysmith engaged in wide anagrams and shifting of letters to try to make it more coherent and actually ruined it. On the message board where Ed did this, he also never claimed or noticed any words appearing vertically, diagonally or backwards. It was only later that Kite and I noticed these words.

So do they all appear by accident? Maybe. But compared to other proposed solutions which also were not trying to create vertical, diagonal and backwards words, why does the RG have 50% more such words and 300% more Zodiac or Crime related words?

I have no final conclusions, I am asking if this may show that at least the much maligned solution, in its RAW form, is worthy of further study. The recent analysis by QT, backed up in part by Glurk and Doranchak, showed good reasons to consider that Zodiac used more L and LL words than average, and thus may be areason to solve the + as L. The Raw Graysmith does, and further work by QT gives some reasons to consider solving some other symbols in the same way the RG does. IMO the appearance of words like BOMBS, LIST, GAME, TIES, SLASH, etc., is an additional factor to consider that at least parts of the RG may be correct and should be used as a starting point for the excellent computer based solving methods used by Doranchak, Glurk and others.

Initial Results

All the proposed solutions total words, Zodiac/Crime relevant:

Cecil: 60, 4 (FAKE, FAKERS, MAIL, FATE)
Starliper: 39, 4 (TIME, HELP, EVIL, LIED)
Kracik: 39, 3 (DEAL, SINS, LIED)
Trav1st: 42, 5 (SEEN, DATE, SHOT, SECT, HINT)
Lathers: 46, 4 (LANE, TIED, SINS, LASS)
Turkmen: 49, 2 (LIES, FOIL)
Average of Solutions: 45.8, 3.66
Raw Graysmith: 69, 11 (BOMBS, SLASH, TELL, LIES, DUEL, GAME, LIST, BOMB, TIES, BARS, LASS)

The Raw Graysmith has a 50% higher total word count than the average, and a 300% higher Zodiac/Crime Relevant word count.

"Overly hasty conclusions". Wait I have not reached any conclusions yet! I have prefaced all my posts by saying this is an initial look that needs to be studied further.

I’m mainly referring to this conclusion: "It is an outlier for sure." Sounded pretty definite to me.

Where did these solutions come from? You created them? With what parameters and intent? The intent to create vertical and diagonal words?

Yes, they are solutions that I’ve created to demonstrate that "interesting" words easily form by pure coincidence, and to show that the solution samples you are drawing inferences from is not large enough.

I am interested for your take on that. Isn’t 50% usually considered a significant deviation? What is your preliminary explanation? Do you think it merits further study?

RG deviates from the average only when you limit the set of solutions. I just showed four additional solutions that contain many more words than RG. I’m confident that many more of them exist, which would drive the averages up even further. Therefore, there is no reason to believe RG is an outlier.

And while my Zodiac/Crime words designation is by nature somewhat subjective, I tried to fair. Why is the average for the 6 other proposed solutions 3.66 Z/C words while for the RG it is 11? That is 300% more.

Again, it’s because your sample size is low. I provided four more samples that show much higher averages than the RG.

What I claim may be significant is that Graysmith did NOT do this. To my knowledge he never published a raw solution, and in all his books and writings he never claimed "Hey BOMBS and LIST appear diagonally", and to this day has never noticed it or commented on it.

That’s not what’s happening. In the act of finding words that appear from left to right in the decrypted plaintext, additional words are accidentally appearing in other directions. It has nothing to do with intent.

So do they all appear by accident? Maybe.

I think it is more precise to say it this way: There is no reason to believe the words are appearing by design.

OK now I see a little more clearly where you are coming from.

In this study of 6 other solutions and the RG is where I accurately stated the RG was an "outlier". In that limited group it is an outlier.

How many solutions did you create to get those 4 that have more words than the RG?

Gee, too bad with all those fancy computer programs you can’t find a solution. You must be doing something wrong. I think you make it too complicated.