That is what I am saying. "Price", that is $/BTC, leads difficulty. It is tightly coupled to operation cost.
A little further analysis shows that we are already there. When you total costs for power, hosting, depreciation, etc. you will find that we have substantially, already reached equilibrium. Any changes to this will be more or less incremental. The ratios will be very solid until we see a quantum leap in efficiencies.
I disagree. You analyzed only a short period of time, a time far from equilibrium. Long-term price/difficulty ratios are trending down:http://bitcoin.atspace.com/income.html
and there is still a lot of going on and the equilibrium is not yet achieved.
The trend was down due to increases of efficiency in CPU mining and introducing of GPU mining. But until recently, mining was oligopolistic. Only tech-savvy people were able to mine with GPUs but it has become more "democratic". Pools which are a recent invention also changed the dynamic quite a bit. These processes are not over. Moreover, mining is still too profitable for anybody to drop out. Mining income for ATI cards is still a few times the electricity prices. Including equipment depreciation does not change it. And there is still a possible a dynamic of concentration of mining in countries with cheap electricity. $/BTC/difficulty ratios that are unprofitable for countries with expensive electricity will still be fairly profitable for those with cheap electricity.
And there is still possibility of botnets or others with "free" electricity and the recent "mystery miner" that changed current difficulty substantially might have been such an example.
I think the difficulty/BTC price ratio will be still trending up (except for the current round where the "mystery miner" artificially pumped the difficulty). Much more slowly than during recent mining gold rush but still respectable 10%+ per 2016 blocks until probably around 150,000-200,000, where it will slow further.