You are discussing a classification model with a product team that needs to understand how model quality changes based on what kinds of mistakes you are willing to accept. They want a simple way to reason about when to favor catching more positives versus being more selective.
How would you explain the trade-off between precision and recall to a product team?
