The medieval 7-sided dice and the Voynich Manuscript Experiment

[Stellarwest]

Random Voynich Vords

Vid explanation:
https://youtu.be/ndKbX0P09CE

After a year of sitting back and thinking about the anagram Morse code translation for the Voynich Manuscript; I have come back with another answer, as an explanation for the Voynich Manuscript text's behavior. I researched a little and found out during the medieval age that 7-sided dice were used for:

They were used with a form of 7-player Backgammon and a variation on Chess which was played on a board 10 squares on each side and extra pieces, the dice were used to determine which Chess piece you could move.

d7ms.jpg (29.76 KiB) Viewed 20395 times

https://historicgames.com/xcart/seven-sided-dice.html

! However, what if someone in the early 15th century used the dice to randomly seed the Voynich text via dice. Here below are the stats for the voynich glyphs and to the right is a roll of 3 seven sided dice for a count of 100 times and look how close the relationship is for a frequency to the voynich glyphs. Wow thats close so I thought to myself this is why the voynich manuscript cannot be decoded it is just gibberish created by rolling the dice.

Here is an excellent program I created in python which accurately displays voynich vords from the random stats of the 7-sided dice and you will be amazed.
https://drive.google.com/file/d/18wn3Sd ... 1Z3oX/view

click the download down arrow in upper right.
1) unzip all files to a root directory
c:\voynich

2) Right click voynich font, open font viewer and install voynich font
3) run Random Voynich Vords.exe


[Voynich Glyph Frequency | 7_Sided-Dice Frequency fix 9/21/2020
Statistics from Voynichese
o , 21895, 14.00% 13%
9 , 16602, 10.50% 11%
c , 14319, 09.10% 10%
a , 13614, 08.10% 08%
8 , 12505, 08.00% 08%
h , 10845, 07.00% 07%
1 , 10674, 07.00% 07%
e , 9957, 06.30% 06%
y , 7170, 05.00% 06%
i , 7053, 05.00% 05%
k , 6872, 04.40% 05%
N , 6140, 04.00% 04%
s , 5576, 03.50% 03%
4 , 5421, 03.40% 02%
3 , 4464, 03.00% 02%
g , 1620, 01.00% 01%
K , 945, 01.00% 01%
H , 907, 01.00% 01%
f , 143, 00.00% 01%
J, 110, 00.00% 01%
F , 47, 00.00% 01%
__________________________________
156,879 Total

A random frequency of 3 seven sided medieval dice rolled 100 times,
which closely resembles the frequency of the "Voynich Glyph Frequency")



Sum Frequency Relative Frequency Experimental Probability
-----------------------------------------------------------------------
3 0.000000000000000 0.000000000000000 0.000000000000000 %
4 4.000000000000000 4.000000000000000 0.040000000000000 %
5 2.000000000000000 2.000000000000000 0.020000000000000 %
6 2.000000000000000 2.000000000000000 0.020000000000000 %
7 2.000000000000000 2.000000000000000 0.020000000000000 %
8 6.000000000000000 6.000000000000000 0.060000000000000 %
9 5.000000000000000 5.000000000000000 0.050000000000000 %
10 8.000000000000000 8.000000000000000 0.080000000000000 %
11 11.000000000000000 11.000000000000000 0.110000000000000 %
12 7.000000000000000 7.000000000000001 0.070000000000000 %
13 13.000000000000000 13.000000000000000 0.130000000000000 %
14 7.000000000000000 7.000000000000001 0.070000000000000 %
15 10.000000000000000 10.000000000000000 0.100000000000000 %
16 8.000000000000000 8.000000000000000 0.080000000000000 %
17 6.000000000000000 6.000000000000000 0.060000000000000 %
18 5.000000000000000 5.000000000000000 0.050000000000000 %
19 3.000000000000000 3.000000000000000 0.030000000000000 %
20 1.000000000000000 1.000000000000000 0.010000000000000 %
21 0.000000000000000 0.000000000000000 0.000000000000000 %]

[R. Sale]

Your dice calculation is in need of a serious revision. One hundred rolls is not enough.
The probability of getting three 1's is small, but not zero; similarly three 7's will make 21.
You need to make a theoretical computation.
You have three dice: A, B & C. There is only one way to total 3: 1, 1, 1.
To make 4 the combinations are: 1, 1, 2; or 1, 2, 1; or 2, 1, 1. So, three different ways to make 4.

To make 5 the combinations are: 1, 1, 3; 1,3,1 and 3, 1, 1. PLUS: 2, 2, 1; 2, 1, 2 and 1, 2, 2
So six different ways to make 5.

As you proceed, this calculation will rise and then fall like a bell curve.
When you have calculated the total number of possibilities for all numbers, then you can determine the percentages.

[Stellarwest]

@Rsale,

I don't doubt that three 1's or 7's should turn up more than zero, yet maybe the number just did not register for a 100 rolls within the confines of a computer program. Believe me if someone did roll the dice manually they would probably shoot for a 100 if they were going for a glyph frequency count to simulate a language frequency count. Then simply tape the glyphs according to their percentages numbers on the dice.

A person has to wonder what a die, would be doing in the Voynich Manuscript? It seems so out of place that a single dice is in it amongst plants, astronomy, astrology and nude nymphs.

Single Die f102v2.JPG (28.05 KiB) Viewed 20370 times

Also online at Wikipedia for the Voynich Manuscript it states that the manuscript is suggested to be made in Italy and well the frequencies are for 21 which is the Italian letter total for the alphabet yet this could be just a setup for the gibberish mask.

[proto57]

Hi Stellar...

That cube is certainly unusual, and also the way it is drawn in perspective. There have been a few suggestions as to why it is there... I can't off the top of my head remember any of them!

As for your idea that the text could have been generated in some random way like this, I could see that myself. Gordon Rugg famously suggested it with his system of word part lists, chosen with Carden Grille-like overlays.

The use of some sort of "machine" to generate the text is not implausible, I think.

Rich.

[R. Sale]

I won't comment on the merits of this suggestion. Validity cannot be determined by comparison to faulty statistical information.

According to my rough calculation, there are 275 possible combinations with three seven-sided dice. So a sample of only 100 roles is totally inadequate to create a valid statistical distribution. If statistical data is the basis for this comparison, then the statistical data-base should at least be accurate.

[johncomegys]

Hi Tom,
Using the 7 sided die to generate various VMs letters is intriguing. I also enjoyed the video. I ordered one from MacGregor Games.
I have three questions.
1) How do I open the program you referred to? I was not able to open the program that begins https://drive.google.com/file/d/18wn3Sd ... 1Z3oX/view
2) What transcription alphabet are you using? You never say. It looks like Frogguy at first glance. It certainly is not EVA which most of us use.
3) Have you compared your statistics to those of Dr. Andreas Schinner in his Cryptologia article of 2007 The Voynich Manuscript: Evidence of the hoax Hypothesis? He describes a stochastic process called "Browning Walk with replacement" that has some features in common with your 7 sided dice.
Cheers,
John Comegys

[Stellarwest]

Hi Tom,
Using the 7 sided die to generate various VMs letters is intriguing. I also enjoyed the video. I ordered one from MacGregor Games.
I have three questions.
1) How do I open the program you referred to? I was not able to open the program that begins https://drive.google.com/file/d/18wn3Sd ... 1Z3oX/view
2) What transcription alphabet are you using? You never say. It looks like Frogguy at first glance. It certainly is not EVA which most of us use.
3) Have you compared your statistics to those of Dr. Andreas Schinner in his Cryptologia article of 2007 The Voynich Manuscript: Evidence of the hoax Hypothesis? He describes a stochastic process called "Browning Walk with replacement" that has some features in common with your 7 sided dice.
Cheers,
John Comegys

https://drive.google.com/file/d/18wn3Sd ... 1Z3oX/view
@ John Comegys,

Thanks John however, remember I used (three) 7-sided dice, so you should order two more 7-sided dice if your going to mimic the voynich stats the way I did. Simply roll 3 dice and add up that number, if you roll the dice a 100 times that number will come up to equal a certain percentage similar to one of the voynich fonts.

1) John did you download the file? Simply click on the link above and go over to the down arrow and download the whole file. The file should be zipped from my rar package so you might have to unrar it or un zip the file to the c:\voynich directory of course first make the c:\voynich directory.
Then right click the voynich font to open font viewer than install font. From the c:\voynich directory run the Random Voynich Vords.exe

2) The font is the Voynich 123

3) Thanks for the heads up I'll have to look at Dr. Andreas Schinner stats on the Random walks with the Voynich Manuscript.

[Stellarwest]

@John

I found a link for the Voynich 123 font even though I have it packed in my zip file. The font was made by Glen Glaston.

Family
Voynich
Description
Based on the Voynich 101 Truetype created by Glenn Claston to transcribe the Voynich manuscript, updated and corrected for UTF-8 by William Porquet

https://fontlibrary.org/en/font/voynich

I read up a little bit on Schinner's report regarding the Voynich Manuscript as a Hoax. I believe the main thrust of his paper suggested that some of the Voynich's text words repeated geometrically (i.e. if you were using dice to make a vord than copy it to the manuscript) to a serious degree more often than normal languages do statistically. To be precise this is unacceptable for a normal language for words in a range of consistent distances and the languages that were compared to the Voynich were Mandarin, Latin, English and German.

Anyone who thinks they can decode the VMS to a language should obviously read the Schinner pdf. The graphs in Schinner's pdf make a great deal of sense and show how far off the VMS strays from a natural language. Also if the VMS was encrypted with the methods of the day then it would have long been decoded, for then there were not elaborate encryption schemes available. Again if it were a made up language the repetition of words still point to a random walk mechanism. So if the VMS was created through the use of a Table and Grille or dice as the algorithm to setup the randomness for the Voynich's text behavior then this makes more sense than anything I can think of so far in my journey with the Voynich Manuscript to date. I do believe now it is a 15th century document, sorry Rich.

Also my Dice theory seems most applicable as the mechanism for random vords, however I'm still not clear on how the dice roller would have managed to build the vords to add to the document. One idea is that the dice roller would have sets of dice for vords. After a person runs an average for a 100 rolls with three 7-sided dice and figures the frequency for each letter in the invented language, that is to setup for the dice rolls. Now if you want to roll a two letter vord then you would need six 7-sided dice and roll them all at once, then the first letter would be a from the first three dice (i.e. the first 3 dice numbers would come from the highest average from the 100 rolls) then the next three dice would be the other letter. So as you can see if you wanted to roll a three letter vord a person would increase the dice to nine and repeat the same procedure and remember to use the highest average to lowest per 3 dice letter.

While writing this one idea jumped out at me regarding the manuscript as another reason why the VMS is random it lacks a varying amount of one and two token vords that would make the manuscript to be a normal text.

Schinner Pdf
https://drive.google.com/file/d/1wrtfy2 ... sp=sharing

[Stellarwest]

The repeats are to tight and close in distance geometrically and Schinner was not referring totally to words in the Voynich that repeat right next to each other. Languages are found in Schinner's Pdf behave with repeated words on average behave at longer distances. The Voynich Manuscript was found to be a mechanized random gibberish text purposely made this way to garnish attention and value.

[johncomegys]

Tom, yes I got it downloaded. All of it. now I must find time to read it, or at least skim it and read the juicy parts. Thanks 1,00,000 John