Well what I'm claiming is not that averages can be used to arrive at all truth, or to observe all things. Many things however are distributions and averages can

Well what I'm claiming is not that averages can be used to arrive at all truth, or to observe all things. Many things however are distributions and averages can be accurately used to characterize them in a way that works. In that example, one might say that the mean viewpoint is correct within the error bars that encompass the two observations, which could be quite large if they are opposite observations.