Even if they started mining today... The break-even point is ~266.3 days. The big problem is people just disregard the average difficulty increase per day. You can see that effect on the "...considering reduction" rows. And this is without taking into account imminent stuff like this: http://thegenesisblock.com/latest-shipment-of-avalon-asics-could-increase-network-hashrate-by-500/
http://www.coinish.com/calc/#
If you were just starting mining with new asic hardware now, what percentage of resulting mined bitcoins should you re-invest in the whole operation to sustain profitability to offset the affects of difficulty increase, and maybe make a bit more headway?
Let's assume it was something like 70% (we can say power costs is part of this). Then wouldn't that be like saying the 30% which you don't re-invest are like dividends you pay to yourself - other than you own the hardware in this case. So, if you took in 30%, what MH value might that equate to per bitcoin invested? Then how does THAT number compare to $15/100MH? And consider that the portion of the hash rate you don't receive as dividends may at least in some significant part going to increase the hashing power of the system.
I'm simply presenting a perspective.
Caveats: I'm not sure what actual numbers would be more correct. Difficulty levels might increase in ways that can't be matched by any plan (other than dumping in new money from outside the system). And it's probably not something to overlook that you don't own the hardware.