Interesting you bring it up. A team at my school is working on an artificial intelligence project which has people answer questions about various games they have played.
From that they're developing a system which analyzes the different aspects of the game against the scores that they average from the press, to be able to predict the score games will get by simply answering questions such as "Does it have multiplayer?" and "In seconds, how long do you spend in a menu before beginning the game?"
I like the idea, but without having hundreds of inputs to develop the neural network, I wonder how well it will work out. I think I'd do good on it:
It's Mario- +2
It involves waggle - -10
No classic controller support - -5
Final score: -13 F
