Interesting.

I beat the poor slob with 'emergency exit'.

But exactly what will this program achieve over time? It says at the end:

It does not matter if our answers disagree, as over time the game will change its answers to reflect common knowledge.

I'm not sure that the program will become more accurate by sorting and indexing the 'common knowledge' contained in our 'yesses' and 'no's'.

I am not familiar with the sort of software commonly used for this sort of thing. However, I am familiar with most of the logical operations it performs, as well as the math (Ven diagrams etcetera) behind it. But I am all too familiar with the highly creative logical operations performed by some, if not most of my fellow-humans. Will the program be able to correct for the x-percentage of unclarity, divergent experience as well as sheer stupidity in the answers it gets?

Example. According to the program, another respondent has answered 'probably' to the question if emergency exits are used by basketball players. Well yeah, they probably are, but not by basketball players as such, i.e. not as part of the game. Since I assumed that was the point of the program's question, my answer had been 'no, not used by basketball players'.

I would love to see some analysis by someone smarter and more computer literate.