The work was done by an artist, not a data scientist or statistician. They fail in the boing-boing article to mention that other rats in the study had as low as

The work was done by an artist, not a data scientist or statistician. They fail in the boing-boing article to mention that other rats in the study had as low as 40% accuracy. There is also some weird breeding they tried to do for generational optimization, but even there the spread was 46.9% - 53.5% Seems to me the sample size is too small to be meaningful (and the results I gather is that any given rat was not much better than random chance... am I missing something?) http://www.artmarcovici.com/rat-traders http://www.artmarcovici.com/rat-traders