We aren't talking about electricity costs but capital outlays.
But you make the decision in the same way. First determine if one can get more Bitcoin by purchasing it off the exchanges than by purchasing mining equipment and mining with it. Would you buy a GPU to mine bitcoin with free electricity today? If not, how did you come to that conclusion?
All of this projection or whatever you want to call it is future prediction.
A projection is determining the outcome based on a fixed set of criteria. A prediction is determining the outcome for all criteria.
Projections are useful because you can determine the set of input values required in order for something to be profitable.
For instance,the following statement is a projection: "If you get your product at 80M difficulty and difficulty adjusts up at an average rate of 9% over the next 8 months, you will not earn a return on your capital investment."
The following statement is a prediction: "You will not earn a return on your investment".
We don't know what the rate of decay will be when approaching the lower limits of fiat costs of chips or a sudden large drop in exchange rates.
One can project the outcome of various rates of decay quite easily using calculators. One might project that a venture would be profitable at a 4% rate of decay but unprofitable at a higher rate over a 6 month period. This is useful information. Of course BFL might go bankrupt tomorrow, KNC turns out to be a scam, Bitfury founders are killed and eaten by Edward Snowden (he is in Russia right now hunting them!), and ASICMiner's facility burns to the ground. Then anyone who already has an ASIC miner will mint money. But those are low probability events.
If you have to rely on a rise in exchange rates to bail you out, your mining is not profitable.
You also forgot to bold the word "NOW" after "Obviously it's a bad deal" in my previous quote. Everything is different with benefit of hindsight and that was my point in response to someone trying to compare deals form a year ago (BFL) vs 9 months ago (Avalon).
I am not talking about months ago (However, a lot of people did post 9 months ago that BFL would not deliver what they promised when they promised it.)
I am talking about right now. I didn't need to bold the "now" because the statement "Obviously it's a bad deal" is in the present tense. People ordering equipment from BFL now must rely on low probability events in order to be profitable. Worse, they did their math wrong in their profit projections and believe that they do not need to rely on low probability events.
