interesting. if i'm reading this right, it's been more or less stable for a year now. that's ...actually really interesting, when you think of all that's happened with the fpgas and GPU's and such. i'm actually quite surprised by this.
The chart against that data shows this visually and you are correct -- relatively flat (log scale).
-
http://docs.google.com/spreadsheet/ccc?key=0AmcTCtjBoRWUdHVRMHpqWUJValI1RlZiaEtCT1RrQmc#gid=1This occurs because mining capacity lags price. The price wasn't high long enough for capacity to catch up to it, so even after the price came down, mining capacity didn't drop at a proportionate degree. And then when the price dropped to levels where almost nobody was mining profitably, many miners kept their rigs running at a loss -- partly due to the sunk cost fallacy but for other reasons as well.
Keep in mind that in less than six months from now the block reward drops to 25 BTC. Some mining operators paying average or above-average electric utility rates (e.g., $0.15 per kWh or higher) know this drop is coming and realize that GPU mining will likely no longer offer a chance for profits (unless the BTC/USD rises much faster than the difficulty by then).
But with the backlog in the FPGA order books, and a hearty appetite yet for new and used GPUs by those whose electric costs are lower, it is likely the difficulty level will be going up, where the rise on the chart is noticeable.