As I said ... you will not get ROI on any SHA256 miner.
That depends on your assumptions.
If you assume non-stop 20% difficulty increases and that BTC prices are stuck at 450, then you are 100% correct.
If you assume 10% difficulty increases and BTC prices of 500, you can ROI on the Spondoolies (at least at the GB price, which expires today) and probably the S2 and Dragons as well.
However, difficulty seems to be at a plateau (next estimated difficulty jump is only a little over 9%), which makes sense. Especially when you consider that we appear to be hitting a hard stop on Gh/watt per dollar. E.g., the hashrate increases that are coming online now in 20nm and 28nm chips cannot be improved on much as those chips turned out to be very difficult to make because of heat issues. Absent a new technology or groundbreaking solution to heat dissipation, there probably won't be a 15 or 10nm process. If I am right about that, then what you should see is 20 and 28nm products replacing less efficient ones, and upping the hashrate, but long term that should be a more linear growth rather than the constant near exponential growth of hash rate that you saw when BTC moved from GPU -> FPGA -> 60nm -> 28nm
As far as BTC price, who can say. I run my numbers at 500 worst case and 700 best case. I believe the current drop to below 500 is going to be short lived and is the result of a number of items of bad news that all hit at once (mt gox, chinese government decision, shutdown of silk road last year and maybe also amazon's announcement that they would not move to BTC). Also in my consideration here is that BTC price is determined in part by miner orders which for many miners must be placed in BTC. When lots high end new miners hit the market, esp on pre-orders, that creates demand for coin and pushes prices up... when miners are shipped prices go back down. I believe the 20-28nm process miners coming out will be generating a lot more sales as folks replace miners that can no longer operate property. I could be wrong of course, but I think 500 as a low end long term (1 year out) is reasonable.