You're missing the entire concept of a test - one keeps verifiable track of all "warnings" for a test period to see how much better than chance is the performance.

It's not a surprise no scientists nor governments are reading your writings.