Ok, so it's more that the AI actually realizes they are negative and would choose positive ones first if possible? That would make too much sense though. Thanks for the reply DG.