Bitcoin Forum

Newbie chodpaba has discovered a strong correlation between lagged btc price and difficulty: http://bitcointalk.org/index.php?topic=4339.0 (http://bitcointalk.org/index.php?topic=4339.0). It predicts that difficulty should stay level or decrease slightly for the next month. We'll see!

Its' too bad that http://BitcoinSportsBook.com doesn't offer a current way to wager on that.  I'ld (hypothetically) put my money on the over, if I could.  Way over.

Hm, betting on the devaluation of the currency sounds like a sure way to make money, if you loose the rest of our money is worth more, and if you win you gain more money; though it would be important to evaluate if the wins and losses on both ends do balance out positivelly....

It would be very surprising (and easy to make money on mining) if difficulty (mining effort) did not follow the price.

Yes.  Except that because difficulty isn't high enough yet to discourage (a significant number of) existing miners, there will still be increases from:

• New miners
  - How much of the press in non-english speaking countries have not had any mention of bitcoin yet.
  - Equipment ordered weeks ago finally coming online
• Existing miners adding more hardware
• Increases in efficiency -- either with the miner and/or in the pool servers

I have my own independent estimate of the whole network's GHash/Second based on slush's posted GHash/Second and the percentage of all bitcoin blocks found by slush's pool. You can see my estimate if you install my Greasemonkey script and view slush's stats page.

For the first time since I started watching the total network number, my estimate fell below 400 GHash/second today. I assume this is a response to the recent drop in bitcoin prices.

Then I realized something. It is extremely easy to remove mining hardware from the network, but much harder to get new mining hardware to add to the network. I believe the two-week lag we see in the correlation between bitcoin prices and network hash rate only applies to increases in bitcoin prices, since it takes time to get new hardware bought and set up. I would expect that a decrease in bitcoin prices will result in a decrease in network hash rate almost immediately, and I believe we have just seen this happen.

Good observation. There is another wrinkle. Mining has a fixed and a variable cost. When deciding whether to join you consider both, but to leave you ought only consider the variable (electricity). I guess you can sell your hardware, but probably slowly and/or at a loss. This should have a dampening effect on miners leaving.

We also might see that a price increase after a drop brings up the rate more quickly than the original climb did because there are now miners who were turned off, but are ready to go any time.

I believe there are a few other things that miners take into consideration.  Even once the initial equipment has been paid for, there is an ongoing maintenance cost.  Most miners are buying consumer-grade power supplies, fans, etc.  These will begin to fail after a few months of 24/7 overclocking duty.  Then there is the noise if you don't have your equipment located in a server room or closet/cellar somewhere.  There is also the problem of cashing out or exchanging your BTC for local currency.  I think a lot of people got into mining without thinking everything through, and some are dropping out due to the expense/hassle/inconvenience.  We will lose a few more when summer comes to the northern hemisphere.  It is already 27C outdoors here.

Last time difficulty was adjusted, we saw a sharp increase. Next time, it will probably decrease, right? Could that trend continue? Maybe some people switched off their miners right now (difficulty too high) and will switch it on after the next decrease. If this trend builds up, the low difficulty phases would eventually become very short, and BTC generation would depend mainly on chance. After these short periods, where everybody turned on their miners, periods with very high difficulty would follow where it is more or less unprofitable to mine.

Just some random scenario that came to my mind.

Doesn't difficulty always change with a fixed frequency?


Btw, is there the risk of some "resonance" building up, where more and more people will mines during easy cycles, making difficulty increaae, then stop mining during hard cycles, making difficulty decrease?

Not sure what this is referring to, but difficulty changes every 2016 blocks. The time this takes depends on the ratio between mining rate and difficulty.

In theory yes. I don't think this will be a problem in practice. People who have already invested in the hardware will not stop mining even after a significant increase, and people who haven't won't go on a shopping spree after a momentary decrease.
If this becomes a problem, the protocol can be patched to fix it. This is basically a control theory problem.

I don't mean buying and seeling hardware, just not turning it on during the periods you aren't gonna have profit to cut the losses(sp?)

I know. The point is that the profit margin will always be large enough to justify the investment of capital (unless the supply of gamers who leverage existing hardware overwhelms the demand for bitcoins). So if, say, the revenues are twice the cost of electricity, the difficulty needs to increase X2 before anyone shuts down their mining. And the fact that people use different hardware with varying ratios just improves stability.