Presenting a very old game, entitled: Man vs. Machine. Or, why Man is not a Particularly Good Source of Entropy.
Stanislav, have you tried playing Rock-Paper-Scissors against the NYTimes computer?
If youre able to “win” this entropy game, you are still losing, since the computer is able to reliably predict what you are going to choose, then play the *opposite*.
Took a run at this, using the following strategy: would periodically alt-tab over from my work and hit a 1 or a 0, hopefully thereby subverting a lot of the cognitive machinery that makes us bad randomizers.
Interesting result: machine would always Pass. Yet I don’t see anything time-dependent int he source. Any idea what gives?
I lost 461-491 after 1000 trials, statistical significance 1 sigma. I’m definitely not impressed.
@AI: Give it some time. It will start guessing when it has enough samples.
Dear Itai Bar-Natan,
If you were expecting magic, I’m afraid I must disappoint you. It’s only a ‘Markov chain.’
Interesting. After 40 moves I was leading 8-4; with a hundred random bits from random.org I was losing 28-31; I took your advice not to spend all day on it.
I’m curious about the scoring, though. From Shay’s comment it seems that “real” score should be derived from something like how often one can make the computer pass, or with alternating rounds trying to match and trying to mismatch.
Mail (will not be published) (required)
XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong> <pre lang="" line="" escaped="" highlight="">