Okay, forget the math. In this case, the “analytics” can be explained to coaches in terms of a simple logical decision tree. Under the presumption that you are going to go for the chance to win, you MUST complete a “risky” 2-pt conversion on at least one of the 2 (presumed) TDs that you score.
You are far better off knowing the result of the risky decision after the first TD, so that you can use that knowledge when deciding what to do after the second TD. You made the 2PC? Kick the PAT next time for the high-probability win. You missed that first 2PC? Well, try it again for the chance of at least a tie.
If, on the other hand, you put off the risky call until after your second TD, your first decision (to kick the PAT] was made blind to the outcome of the second play. You make the PAT then miss the 2PC? You are hosed, because you can’t go back in time and exchange the PAT call for a 2P—.you just lost the game, pal. You make the 2PC? Lucky you, you won the game despite painting yourself into a corner.