Different seasons have different sample sizes; weighting by attempts gives appropriate influence to each season.