What you're missing in wanting to wait to go for 2 until you absolutely have to is that in most of those cases, you are totally screwed if you don't convert. Additionally, looking back and saying, "If they wouldn't have gone for 2 'early,' and missed, they might have cost themselves the game," is results-oriented thinking.

Let's look at a hypothetical, yet common situation:

Packers are losing by 15 and score a TD around the beginning of the 4th quarter. Deficit is now 9. Many coaches consider this "too early" and decide to kick the XP to go down 8, rather than trying to make it an 7-point deficit with the 2-point conversion. After all, if they miss, it's a two score game, but if they kick the XP, they can still tie later with the 2-point conversion. 4th quarter proceeds, deficit remains at 8 and the Packers finally score a TD with 15 seconds remaining. They go for 2, miss, and now there's not enough time left, barring a miracle, to have any chance to win the game.

If they go for 2 and miss it on the preceding TD, yeah it stinks that they are down 9 instead of 7 or 8, but now there is enough time left to game plan for multiple scores.

Bottom line: if you know (or believe) you are going to need a 2-point conversion at some point, you want to go for it earlier so that you have time to make up that extra score should you fail, rather than waiting until the eleventh hour when it's do-or-die.

In the Packers' situation yesterday, they should've gone for 2 with 11+ minutes left to try to tie up the game. Realistically, there's little difference between being down 2 and being down 1 in that situation. Sure, you are in worse shape if the Bears score a TD after you miss the conversion, but you can't work on the assumption that you're going to give up another TD at that point in the game. There's little enough time left that no more scores for either team is a distinct possibility, as is just a FG by the Bears. Obviously, a Bears TD is really bad, but that's worst-case scenario. The other, very possible scenarios that don't involve the Packers scoring first hurt your team in a way that could have been prevented via the 2-point conversion (and I'm guessing that their combined probabilities was greater than a Bears TD).