Handling 0% Accuracy for 0 Attempts.

I'm working on my first side project that dips my toes into data science. Specifically, I'm looking at UFC fights and outcomes. To determine fighter's win chances, I'm looking at a few things like strikes per minute, accuracy, takes downs. My issue is some fighters never attempt a take down. So a fighter with 0 takedown attempts has a 0% take down accuracy in the data. This negatively impacts them on the backend because a bad take down accuracy isn't …

accuracy data data science datascience issue per project science strikes