Thursday, July 17, 2003

 

Question of the day:

Given a sequence of N bits, what metric would give you something akin to "how unlikely it is that the sequence resulted from random (p=.5) generation"?

I know there are measures of the 'complexity' of sequences (Kolmogorov complexity, for one). I don't really know how they work, but I know they exist. But they don't really seem like they answer the specific question.

My curiosity was piqued by this sequence: (The history of the MLB All-Star game)

0000001110000001011101111111111101111111110111100011011110000100010101000

Yes, I know this is kind of a tricky (possibly badly formed!) question. :-)

I welcome any and all input.

posted by Miles 3:48 PM