Randomness isn't actually required.

There are no indications what is being tested so the tests might as well be a black box.

Moreover, the description doesn't describe what is actually a noise. Is it the string length? The amount of 1s? The amount of groups in the string? And what about tiebreak?

hi,

• all your patterns lists should be sets (in the tests)
• use for key,pattern in dct.items():
• you should actually do a statistic check rather than use a threshold to test for randomness (but well...)
• is that really necessary to have more than 1000 tests?
• at most by (Y) vibrations in a zone -> (Y)?? :o And you should talk all way long of noise, not vibrations: you'll lose some beginners along the way, mixing the epxressions.

(I'm honestly not convinced by the background and its link with the actual task. Mostly, I've been asking myself all way long "the f... why" I'm implementing all of those methods x)
The main problem being that you ask to us for studying the noise of inputs we generate ourselves (apparently. It's not what is done in the tests, but that's the general feeling the current setup gave to me).)

Some calls of assert_equals have arguments in the wrong order.