Well, it's up for debate in a lot of ways IMO. Here is a thread on Reddit that is somewhat helpful:
http://www.reddit.com/r/Bitcoin/comments/iduxz/new_miners_its_not_profitable_to_build_a_rig/I think the aspect of interpretation with calculators is they feel hash moves in step with exchange price. From what I understand, that's not necessarily true, but alas, I am no expert. What we do have background and intel on is difficulty, power costs, and equipment costs. What we don't is where that price is going to go.
AFAIK, difficulty and exchange price are independent. While they may move close in some respects, a black swan event could push price higher much faster than difficulty. If that was the case, being able to mine only 1 btc a week or something paltry like that, at some point, would then still be cost effective. Which, then, buying equipment to mine would be a "today" play on bitcoin displacing other currencies in the world soon. It being ASIC, and not some future tech around the corner, is also showing that you think said displacement is going to happen within a, I dunno, year or so. Depending on what your interpretation of your analysis may be.