Andome

Well, the guessing probability is the logarithm of the min-entropy, not the Shannon entropy. As for Shannon entropy, it's impossible to get SE with our current code generation because there will always be a small difference from theoretical limit. When the length of our codes get bigger the Shannon entropy/uniformly distribution model breaks down and the Guesswork time isn't as long as we once thought.

Pride goes before destruction, and a haughty spirit before a fall.