The analytics are simple on the face of it. If a team's conversion rate is 50% or better (actually about 48% or better), that team should always go for two, as that will lead to more points over time (and therefore wins) than by going for 1, especially given the ~94% success rate for XPs since they moved it back (of course those are avg stats and some teams have been 100% on XPs). But coaches take a long time to change behaviors and are much more averse to taking chances that are generally considered too risky, but I expect that will change over time, like it has, to some extent, for going for it on 4th down.