The hard part about pre-ordering next generation hardware is to predict a logical expectation of A) what will be the total network power (hashrate) be by the time your miner comes online which influences directly B) what will be the difficulty value at that time.
If you assume linear growth in the network vs exponential for example you will arrive at different conclusions. That is the real hard part.
At todays difficulty (now 908,350,862) a 2 THps miner at $1000/btc can expect to break-even in 5-6 days.
If looking backwards is any indication, the difficulty has increased by 945% in the past 90 days so if that is the same as the next 90 days and BTC values are fixed at $1000/btc you can expect to break even in 47 days.
IF you think that the next 90 days will grow faster (say 1.5x as fast) as the previous 90 days then you can expect a breakeven in 75 days.
What if your miner comes in 4 months from now instead of 3? What if a whole lot of people are buying one so that when they all come on line simultaneously the difficulty will spike beyond your projections?
And what if the price of BTC falls again from where it is now and stays lower?
A lot of what-ifs but you must know your variables before investing in mining capital.