Topic: Difficulty vs Value, and Bubbles! (Read 1152 times)

Difficulty follows network hashing more or less directly.  But it also follows price, indirectly.  

Miners use formula to determine how profitable mining is.  If price increases, mining becomes more profitable.  If difficulty increases, mining becomes less profitable.  If both increase at the same rate, mining profit remains the same.

If difficulty is increasing faster than price, profit goes down.  Miners with expensive electricity start to lose money or barely break even, and lose interest.  Eventually difficulty drops back down because of these miners leaving.

If price is increasing faster than difficulty, profit for miners goes up.  This encourages small time miners to purchase more hardware, and encourages even miners with expensive electricity to mine, because the profit is so high compensates for the expensive electricity.  Eventually difficulty increases to match the higher price.


Right now though, we are in a bit of a special situation.  By all rights, price is getting way out of line with difficulty, mining is extremely profitable.  But the threat of ASIC mining is stopping most miners from increasing their GPU based mining investments, as the common thought is that difficulty will skyrocket in the next few months until GPU mining is unprofitable for all.  But the endless delays on ASIC delivery has slowed things down a bit, and GPU mining is still very profitable.  Difficulty could increase 10X and I would still make some profit with my GPU mining, for example.


My OP was simply recognizing a pattern in the historical data, nothing more.  Can you explain why value would crash after price exceed difficulty by a certain threshold for a certain length of time?  Or is it just random luck that a crash or correction has occurred after each time price increase exceeded difficulty increase?

It used to be, if price was way above difficulty it would make sense to just run out to the store and buy a mining computer and lots of GPU instead of going to the exchange.

Now it is a matter of waiting for a batch waiting list to make some ASICs available to order some machines that may arrive in half a year.

Difficulty doesn't follow price; it follows network hashing power. Any number of factors could affect network hashing power (including, but not limited to price). But I'm sure you already knew that since you've been mining for the past 12+ years

Then you should post graphs comparing volatility, not absolute values. Right now, you have graphs of absolute values which do a poor job when you try to compare percentages.

So two things should never be compared unless they can be expressed in the same units?  Open your mind a bit and realize I was talking percentages.  A 50% increase in difficulty equates to a 50% increase in price.  If price increases by 60% and difficulty increased by only 40%, price exceeded difficulty.

Stopped reading here. You should think about what you wrote and then delete this topic as it has exactly zero sense in it.

Speaking as a miner since November 2000, I would say that difficulty follows price, with significant lag, but with ASICs as a confounding factor.

Consider: At the time of the run up to the first crash, mining was becoming increasingly profitable as price was going up faster then difficulty. As a response to this, I ordered some more 5870s and increased my hashing speed. Other miners did the same, and difficulty went up fairly rapidly. It takes less than a week to go from ordering hardware to mining Bitcoins.

When the price crashed, the question is different - the mining rigs are sitting there so, as long as they make profit, however small, it makes sense to leave them switched on, maybe with lower voltage/clock setting for better efficiency. Only miners who have unusually expensive electricity are going to turn rigs off when the price moves down, but when GPUs fail they will probably not be replaced. We see a slow decline in mining effort when the price is low.

The situation now is somewhat different because of the ASICs.

Avalon have actually delivered some ASICs, and we can expect more to come. Although the current price and difficulty would, in other circumstances, justify investment in GPU based mining rigs, GPU miners may well conclude that the expected difficulty increase in the near future when Avalon batches 2 and 3 are delivered will destroy their profitability - this is my belief - so I would believe that the recent increases in difficulty are due to ASIC deliveries and maybe some new miners using existing GPUs. I strongly doubt that many miners are buying new GPUs at this time.

This means that the future difficulty change is influenced more by the speed at which Avalon (and other ASIC companies) get their machines into the hands of the miners than by any other factor.

Even a serious increase in price (say another 2 or 3 doublings) are unlikely to cause miners to order GPUs for mining - the risk of being left with a pile of expensive to run hardware at a time of much higher difficulty is too high.

If the price drops... well, Avalon batches 2 and 3 are ordered and paid for, so they will be delivered and used making GPU mining a lot less profitable, so some GPUs would get turned off, but difficulty will not fall much as ASIC mining is so much more cost-effective in term of power use. ASIC miners can still make profit if the price falls by a factor of 10 from where it is now and all the GPU miners are forced to quit.

Conclusion: The relationship between difficulty and price is no longer a reliable guide for anything.

Full disclosure:

Currently mining - 4 5970s, 13 5870s and a 6850, about 9Ghash, power cost about £11/day.
On order - one 4-module Avalon rig, about 80Ghash, power cost maybe £2/day.

coincidence? or... you are observing 2 properties which are both driven by popularity?

But it does!  Look above, each crash was predicted using this method.  If you are speculating that it won't predict future crashes, that very well may be the case, but it perfectly predicted the crashes in the past.

No offense, but a lot of people reading this thread are likely thinking "not again..."

Difficulty follows price such that on average, a miner makes a small profit. This mechanism is slow, so it lags behind on huge rallies or crashes.

As for the difficulty spike: The first commercially functional ASICs were ready just when the difficulty hike started. Take a guess what the reason is: your scale factor is off now because ASICs have a higher hashpower per unit price.

TL;DR: this is not a useful method for price prediction. It doesn't detect peaks, bubbles, rallies, crashes, or anything of the sort.

Even if it was a factor then it is not now. There are so many more bitcoins in circulation than the amount that are "minted" each day, miners don't get to choose the price.

Looking at the chart, apparently they're pegging $1 to 70000 in difficulty.

The charts are simply taken from bitcoinx.  The units are on the axis of the chart.  The units are basically irrelevant though, the importance is the percentage change and the relation. 

Price can never "break" difficulty, they are measured in different units. What is the conversion parameter that you use to convert between the two? Or maybe I am missing something here?

I'm not sure if this has been brought up before, but I noticed this today and thought some others might find it interesting:

Historically, as far as that goes with bitcoin, the first major bubble occurred after price exceeded difficulty for more than a few days.




More recently, our brief correction after hitting $74 last week occurred immediately follow a breakout of the value above difficulty.



However, difficulty adjusted a short time later and the current value per bitcoin is well supported.  We may break out again shortly, if value continues to increase, so it'll be interesting to see if there is a correction, following the pattern, until difficulty adjusts again and keeps everything in line.