"You would prefer we draw names out of a hat?"
Never said that.

"You can't go off record alone."
Never said we should go off of record alone.

"But if you want a model driven approach but don't want the model to be accurate then I guess we just disagree."
I never said that I don't want the model to be accurate.

I am making a point about fairness, about what is fair to include in the selection criteria. I am content with sacrificing some predictive accuracy to ensure that the final rankings reflect only on-field performance for the current season. That said, I'm not even convinced that the final ranking predictive accuracy of such a model would be distinguishable from a model that did leverage prior season data. But what it would do at the very least is give fans and observers more confidence in the selection process.