How should the actual values be ranked? Given that some values have more than 10 decimal places, is it advisable to truncate these values to just 3 decimal places before ranking, or should the ranking process incorporate the entire precision of up to 10 decimal places?
Regarding predictions, in cases where identical ranking results are obtained for the some subjects, should these subjects be assigned the same rank, and should they be additionally labeled to reflect this equality? Or just label them in order and denote they’re tie.