I think you are furthering the point of prediction vs. modeling, right?

Kinda. I just wanted to point of in general why the models are categorize as statistical model vs non statistical models. Certain area of statistic do prediction aka forecasting too (time series). It's just statistical models gives CI for their prediction and ML usually do not give any, most of the time, it's created empirically and have no theory to get a CI.

As for Confidence Rating, I have no idea what it is I've tried Google but I couldn't find much. What field is this?