Machine learning systems can make decisions, but it's important to understand why they make those decisions. One way to explain these decisions is by creating "counterfactual instance explanations," which show alternative scenarios where a person gets the decision they want from the machine learning system. However, just showing these scenarios isn't enough to fully explain the decisions. To provide a satisfactory explanation, both counterfactual instances and causal equations are needed. By combining these two elements, we can better understand and explain how machine learning systems make predictions.