Cherry picking unsuccessful attempts is stupid.

And is no different than criticizing games where a team kicks instead of goes for it and then loses by 3. Except that people don't generally do that. Those games do not stand out, even though the decision to kick it is what lost the game because it wasn't the decision to maximize point totals.

Unless you can show me data showing that conversion rates are not as frequent as the assumed data set, based on all available data, going for it on fourth increases scoring outputs. On a super macro level, a 1.5 to 2 points per game increase in scoring per team compared to 2007-2010 when the analytics started being more widely spread would be about what you'd expect from the strategy change. But obviously there could be other changes that would need to be isolated, such as rules changes and other strategies/player improvement that may make TDs more likely even assuming three played downs. And despite going for it more, teams are actually be making more field goals because kickers are both making more field goals and being asked to kick 50+ field goals at a greater frequency. I don't know if the newest models have accounted for the greater 50+ FG accuracy in determining whether to go for it.