yes, i know how difficulty works
the losses will be fairly consistent throughout the various batches.
i guess that is one opinion, just not mine
Its not just an opion, its math:
https://bitcointalk.org/index.php?topic=283820.0If we assume most of these asic vendors will indeed ship their 28nm according to their promises, 8PH will be added to the network before the end of this year, and another 7+PH is the months thereafter, If you dont believe they will, then neither will your presumed profitable asics ship, so it doesnt really change a thing, just delays everything. That 8PH is a very conservative tally that doesnt include a lot of potentially big players like ActiveMining or next gen chips from either Avalon or AsicMiner and isnt updated for Cointerra's second batch nor 2 new asics announced today. 10PH in december (+whatever delay you expect) and another 10PH in the month after is more realistic IMO.
An added 10PH means a difficulty of ~1.5 billion and doubling that, well, you guessed it, ~3 billion a month later. At those difficulties and current BTC exchange rate, your $14000 2TH miner will generate $2,660 in the first month and $1,330 in the next. Good luck with that ROI.