The point is that stats isn't useful for demonstrating the existence of bias since we know already the bias exists given the deliberate introduction of bias. We

The point is that stats isn't useful for demonstrating the existence of bias since we know already the bias exists given the deliberate introduction of bias. We know the population function is biased and sample data from that function isn't going to change that. The stats in the post is mostly a commentary on the phenomenon of statistical power. It may also be a commentary on how the angle of the bend is non-linearly related to coin flip bias. I did find that bit to be unexpected. Although given the wide sampling distribution due to the low N, this could easily be variance.