Hypothesis Testing: Overview and Tools
Posted: Thu Jun 30, 2016 5:49 pmI posted a new article and tools on the ZodiacKillerCiphers wiki for my “Hypothesis Testing” approach:
http://zodiackillerciphers.com/wiki/ind ... is_Testing
The basic idea is that we don’t yet know how Z340 was constructed, so let’s rule out some encipherment schemes by making a bunch of test ciphers. If we can crack the test ciphers but not Z340, then we can exclude a hypothesis and move on to another one. So far, I’ve generated test ciphers for three different schemes: 1) Homophonic substitution, 2) Columnar transposition + homophonic substitution, and 3) Scytale transposition + homophonic substitution. I will gradually add more whenever I find the time.
Each hypothesis is linked to at least 100 test ciphers. You can get them as a big text file, or you can use an interactive cipher feature explorer tool that I built to visualize Z340-like features in the generated ciphers:
hypothesis 1: http://zodiackillerciphers.com/hypothes ... dex-1.html
hypothesis 2: http://zodiackillerciphers.com/hypothes ... dex-2.html
hypothesis 3: http://zodiackillerciphers.com/hypothes ... dex-3.html
Works best in desktop web browsers - I don’t recommend trying to use it via smart phone web browsers.
Some quick instructions: Use the arrows and the drop down list to move between the different ciphers.
Hover over the different buttons to highlight different features.
(notice that the cipher is re-written to 19 columns to better show the period 19 repeats)
(notice that it shows the full sequence of symbols at the bottom)
The link at the bottom takes you to the text file that contains every cipher for that hypothesis:
Let me know if you find any problems with the tool.
There are so many more hypotheses to explore. Hopefully we can narrow the search to the most probable ones, since there are so many to pick from. One way to narrow the list is to see which schemes have higher rates of production of rare phenomena such as the pivots and periodic repeating bigrams.
http://zodiackillerciphers.com/wiki/ind ... is_Testing
The basic idea is that we don’t yet know how Z340 was constructed, so let’s rule out some encipherment schemes by making a bunch of test ciphers. If we can crack the test ciphers but not Z340, then we can exclude a hypothesis and move on to another one. So far, I’ve generated test ciphers for three different schemes: 1) Homophonic substitution, 2) Columnar transposition + homophonic substitution, and 3) Scytale transposition + homophonic substitution. I will gradually add more whenever I find the time.
Each hypothesis is linked to at least 100 test ciphers. You can get them as a big text file, or you can use an interactive cipher feature explorer tool that I built to visualize Z340-like features in the generated ciphers:
hypothesis 1: http://zodiackillerciphers.com/hypothes ... dex-1.html
hypothesis 2: http://zodiackillerciphers.com/hypothes ... dex-2.html
hypothesis 3: http://zodiackillerciphers.com/hypothes ... dex-3.html
Works best in desktop web browsers - I don’t recommend trying to use it via smart phone web browsers.
Some quick instructions: Use the arrows and the drop down list to move between the different ciphers.
Hover over the different buttons to highlight different features.
(notice that the cipher is re-written to 19 columns to better show the period 19 repeats)
(notice that it shows the full sequence of symbols at the bottom)
The link at the bottom takes you to the text file that contains every cipher for that hypothesis:
Let me know if you find any problems with the tool.
There are so many more hypotheses to explore. Hopefully we can narrow the search to the most probable ones, since there are so many to pick from. One way to narrow the list is to see which schemes have higher rates of production of rare phenomena such as the pivots and periodic repeating bigrams.
