The strength and weight of evidence in backcountry avalanche forecasting