Every such analysis will have some hindsight bias. Still, it’s a great post that shows how to model such behavior. And I agree with the another reply that not deleting an active plan seems like a basic fail safe choice which the post also covered