It’s an FR tradition to not even read the article. You want me to read the thread?
Well, you have me on that.
I explained how you could have taken a random sample of similar cases and tracked their progress. The issue is you might not really “save” anyone...just improve the results marginally.