Input: a CSV group,option,successes,total.
Print the overall success rate of each option (alphabetical): overall: app_routing 78.0%, manual 82.6%; then for each group (alphabetical) weekday: app_routing 93.1%, manual 86.7% (app_routing n=87, manual n=270). If one option wins (has the highest rate) in every group but is not the overall winner, print SIMPSON'S PARADOX: app_routing is better in every group, but manual looks better overall and reason: the options are used in different proportions across groups with different base rates; otherwise no reversal between the overall and the per-group comparison.