There are certain models that may agree with this given the right situations for sure.I get it in that context and agree. Appreciate the explanation.
I do think analytics are great and useful, just think a coach has to only take them as part of the equation.
Personally I think playing for two 7s being down 14 can be a wise decision depending on all the other variables. If your defense is total shit then I guess you gamble on getting a stop and roll the dice. I also don't like the momentum switch that occurs when you fail on a 2.
Like models with PFR use vegas odd to help make better predictions on what would happen. Take for example your team is far better than the other team on paper and Vegas odds reflect that by having you be the heavy favorite. If your team was just unlucky and were down 14 because of fumbles/tipped INTs/Bad reffing it does not mean that your team still is not the far better team. More do if you are only down by 14 and the other team only scored on those unlucky plays by you.
If you think that is the case as a coach then going by the conventional knowledge and just take those extra points its assumed that your team also will find a way to win because well you are the better team.
Flip that in reverse where Vegas thinks you are the heavy underdog and you are down because you are the worst team and those 14pts were put up on you in a clean way then it makes sense to try and go for the 2pt conversion and because you think in conventional situations your team would be the one losing.