This experiment is interesting, but is fairly limited because it is hard to investigate the hop stand space with two data points. For example, I tend to do much longer hop stands. Also, the sample size for the test was small. Specifically, there were only 12 testers so 50% of the testers identifying the odd beer out did not reach statistical significance, but if the 50% of 20 testers did so the results would be statistically significant (IIRC).