 This study examined the performance of infectious disease forecasting models in California counties during the COVID-19 pandemic. It found that the models varied in their accuracy depending on the county and the time period. Additionally, the model's performance was affected by local transmission trends, variant prevalence, and county population size. To improve the model's accuracy, the authors suggest incorporating geographical heterogeneity in model coverage and performance. This would allow for more consistent model reporting and improved model validation, thus strengthening the role of infectious disease forecasting in real-time public health decision-making. This article was authored by Lauren A. White, Ryan McCorvey, David Crowe, and others.