User blog comment:Felza/40% Troll vs 50% Troll/@comment-5266525-20131029153836/@comment-12282825-20131030122423

@Shaw Fujikawa: Actually Sample size is more than just relevent. In statistics, be it used in science or other subjects it is always possible for the data to be skewed. The theory does not target an individual case, but if 40/50 is the average, you may not always experience this. With 10 games, you can easily have 50/50 and with 4 you can easily have 3/1 ratios.