How many months of trading data do you think a trader should collect before deciding whether a strategy has a genuine edge?