To understand situations and make decisions, our minds rely on cognitive coherence, constructing mental images to explain concepts. For instance, we have a mental image of summer as bright and hot. However, we tend to place excessive confidence in these mental images, even when data disagrees. For example, if the weather forecast predicts cool summer weather, we may still dress based on our mental image of a hot summer and end up shivering outside. To make better predictions, we can use reference class forecasting, which involves using specific historical examples to improve accuracy.

For instance, think of what you wore on a cold summer day in the past. Additionally, creating a long-term risk policy that plans for success and failure in forecasting can enhance decision-making. This approach allows us to rely on evidence rather than general mental images. For example, in terms of weather, it could involve bringing a sweater just in case. In this way, we can improve the accuracy of our decisions and avoid overreliance on often faulty mental images