points
It feels like making "Where the tokens land was checked against the table published in Scientific American in 1996 (Illinois Avenue: 3.18% published, 3.19% in my simulation)." a key point of the investigation doesn't see like a human choice