comment: Acting on a correlation is appropriate when there is high confidence that the correlation will recur in the future and the risk of being wrong is low. When developing data science models based on correlated variables, it is important to consider the frequency of the correlation and the clarity of causality, as these factors increase the confidence in the predictive power of the model. Complex models do not always require a full understanding of why they work, but clarity of cause can be critical in certain situations, such as when the risk of acting on the correlation is high. Although correlation can sometimes justify taking action, in scenarios where the consequences of a wrong decision are severe, it's important to grasp the underlying causality of the correlation