Ruby's Greek Thread

[Ruby Novacna]

I can do something similar. 

1. Let’s assume that Voynichese is encrypted Latin. 
   
2. Let’s take f33r. 
   
3. The first words of the page are tarar daiin. 
   
4. The Latin name for the sunflower (according to the translator) is Messier. 
   
5. We can see that there are coincidences: Messier and tarar daiin. 
   
6. We acknowledge that i and n are the same letters and are equal to e. 
   
7. The letters m, i, and r remain undeciphered; 
   
8. in theory, they should correspond to the letters t, d, and ai.
   

What am I getting at? The point is that the entire analysis is not focused on deciphering, but on finding similarities with a specific word.

You've described quite well the method that I think all substitution researchers use in their work.
You're just missing the description of the next step: applying your findings to other words in the manuscript to see how many words conform to your rules—10, 100, 1000, 10,000? (Just kidding, there are fewer than 8,000 words in total.)