Look at simple facts and logical inferences.
If you fail, in either case, you're likely done, because you end up with 0 points and give the ball back to the opponent with 3 minutes or less left in the game, and a 2 score lead . The failure rate in going for it is higher than the failure rate in kicking the FG.
Advantage FG, and that kills the "• The Falcons need to go for it on one of the two drives anyway, and a failure will be demoralizing regardless of when it comes. " argument put forth by Barnwell. This is part of why Oswlek is wrong in claiming the FG% doesn't matter.
If you go for it, you're not guaranteed to stop the clock if you convert. You're more likely to have to run more clock down, while kicking the FG stops the clock immediately. You also are still not guaranteed to score and you're still down two scores with less than 3 minutes left. Nobody knows how that next series of downs plays out, and a penalty of play for a loss could result in just the sort of 4th-and-7 he's talking about
Advantage FG, and this kills his "• Not only was the fourth-and-1 they passed up easier than the fourth-and-7 the Falcons eventually had to try to convert, it was easier than the average situation they would expect to be in on the subsequent drive.", particularly because he seems to ignore that such a "subsequent drive" will be needed whether they score a TD or a FG.
His "• The only way to win the game in regulation is by going for it on the first drive of the two." is meaningless, because you need to not lose in regulation much more more than you need to win in regulation.
Even the best argument he has is a lousy one. His "• They lose if the Patriots score on their next drive." ignores his own, earlier arguments about needing to drive the ball, ending up with 4th downs and the like.
Essentially, what he's trying to do is take a worst case scenario and apply it to the FG, while taking an absolutely best case scenario and applying it to the conversion attempt.