In multi-target situation, why CalcGainGivenWeight fomulas use ThresholdL1 instead of CalcGainGivenWeight or CalcGain?
thank you!