One way to assess the practical accuracy of two different forecasts is to consider what happens in reality when the forecaster makes an extreme forecast. The idea is that decision makers might not change their behavior unless a forecast is extreme. They might risk scheduling baseball or soccer practice if there's only a 50 or 60% chance of rain as per the weather expert. But if there's a high likelihood of rain, they will cancel practice. Whose forecasts are more accurate conditional on their issuance of an extreme forecast