The best thing about (maybe this didn't make it to deltoid) was that they counted degrees of freedom for the t stats based on the number of days the experiment

The best thing about (maybe this didn't make it to deltoid) was that they counted degrees of freedom for the t stats based on the number of days the experiment was running, which was incredibly effective in stopping people launching broken crap.