Sorry then, I guess I missed the whole point. I thought the whole point was a debate of whether teams'/coaches' decisions should be based solely on what the numerical analytics said to do. I only mentioned Pittsburgh's 2 point strategy because it was based on analytics.
So, in a nutshell, and speaking only of the 4th down scenarios in the OP, I'll say again that what I would depends on my assessment of the variables beyond the analytics. I might go for it in 1 situation and not in another (same down distance, point in the games, etc.).
I agree that coaches do tend to the conservative side and probably could benefit being less so, but it depends on the individual coach and team. The gravitating to the conservative side is probably born from the fact that failure gets you fired (unless you're Billy Martin working for Steinbrenner).