It Is Possible to Evaluate Presidential Forecasts in Less than Millennia

There is a long-running debate on evaluating forecasts of rare events such as presidential elections. Recent scholarship has argued that forecasts cannot be evaluated on practical timelines and therefore do not provide reliable information. I argue the opposite: that a variety of statistical methods indicate that election forecasts can be evaluated in a timely manner, and that key forecasts that have survived many election cycles are likely to be informative.