So my seven-coin result yields p < 0.13 – slightly lucky but no evidence against the coin's fairness. The next time I tried it took me 17 attempts. Statistics are not useful unless you predict the result in ad