I know this may involve lots of coding, however if we had 50+ columns, wouldn't the WNH differences be much lower thus would be much easier to predict the outcome with a simple -5+5?
Has this been attempted before and what were the results? Because the idea is not to predict the outcome of 50+ columns but instead would permit us to cover all possibilities of the outcome with a simple -5+5 or even lower. Thus no predicting per say is involved. On the opposite side, only 1 column would give big diff ranges thus harder to predict the outcome.
Would a beta plugin be a possibility for those who want to test with a configurable amount of columns of their choice?