Days Out refers to the forecast horizon, defined as the number of days between the date the forecast was collected and the target date it is predicting. For instance, a forecast collected on June 20 for June 21 is considered “1 day out.”
The further “Days Out” a forecast is, the higher the tolerance for error; however, a 10-degree “Bust” at Day 1 is viewed much more critically than a 10-degree miss at Day 7. Industry professionals search for “Error Growth Curves” to visualize how Absolute Error increases as the lead time moves from Day 1 to Day 10.
Day 1 (tomorrow) is the most critical for end-user decision-making, such as school closures or energy load balancing. Verification systems often prioritize Day 1 and Day 2 accuracy scores because these forecasts have the highest economic impact. Searches for “Day 1 vs. Day 3 accuracy” are common when companies are trying to determine which provider to trust for short-term logistics.
“Day 0” is often referred to as “Nowcasting.” It involves the observation-heavy updates for the current day. While many glossary terms focus on Day 1 and beyond, Day 0 verification is essential for high-impact events like severe thunderstorms or flash floods where the “Days Out” is measured in hours rather than days.