It’s not nonsensical. The line has to be drawn somewhere, however arbitrary it may be.
Maybe you didn't notice that I very carefully set it up to compare two 0.045's with a 0.06 and an 0.01. (I.e. the combined probability is less than two 0.045's.) Or perhaps you've forgotten that a large part of the reason for requiring two trials is not to get the 0.05 squared but to show repeatability. (I really am not meaning to be confrontational, but I do disagree. I *might* agree you have a point if it was an 0.06 and 0.49 - the line does have to exist somewhere.)
Market Data 
Markets 
AI
