Talk:Cognition August2018/@comment-35453425-20180528203135/@comment-35476160-20180530150422

I have some thoughts you guys will probably be able to help me out with so I can understand this all a little more clearly. So greater certainty with higher cardinality is due (in part) to the greater absolute difference in the ratios (ex: 4/10 vs 8/20 there is a difference in blue to orange of 6 and then of 12)? They talk about this as part of the explanation and I think that is where I start confusing myself.

When I look at it this way I can see it both as the clearer difference makes it easier to make an association to “more orange” but also that it’s a clearer difference versus blue so is it the same thing as using a probability? Does that make sense? Is this sort of what you were saying, Mike? Or am I missing something. Because, since accuracy is best accounted for by the probability distance between the two samples, it’s not JUST because of absolute numbers, but couldn’t it be in a way? Again, I think I’m confusing myself the more I think about this.