4 ms·
Since difficulty recalculation is based on number of blocks, not time, would a scenario be possible where hashing power keeps crashing constantly pushing block
by hippich 4y ago
Since difficulty recalculation is based on number of blocks, not time, would a scenario be possible where hashing power keeps crashing constantly pushing block difficulty recalculation in the future and essentially freezing transactions on the network?
- wmf 4y agoI don't think there's any kind of feedback loop where a drop in hashrate would cause further drops. Even if hashrate dropped 90% then you'd have to wait 20 weeks for difficulty to adjust but the network would still function somewhat.
- maxbond 4y agoPurely speculating, but there could be a feedback loop if people observed the hash rate plummeting, lost faith in the network's functioning, and selling en masse. The plummetting price would impact the profitability of miners, who might start pulling capacity from the network, confirming people's fears and encouraging more selling pressure.
- jiggawatts 4y agoI did a back-of-envelope calculation that it is possible for BitCoin to get into this kind of death spiral, but it's not very likely. In the hypothetical scenario that the mining rate drops to just tens of percents of the original, then the utility of BitCoin drops because transactions get both slower and more expensive. This could cause a "run on the bank" where people try to get their money out, but can't transact on the network. If the next hash rate change is far enough out and a panic sets in long enough, then BitCoin's value could plummet as users lose all confidence. This can recover, but if it's juuust bad enough to take out more miners via bankruptcy after the next hash rate adjustment, then it can sustain itself. Miners have fiat expenses and BitCoin revenue. If BitCoin drops too fast, they have to cease operations, otherwise they're literally burning piles of valuable currency to make worthless monopoly money they can't spend on anything. It all depends on how long the big miners can "hold out" while being unprofitable, or more importantly.. how long they're willing to risk losing everything they've made until then.
- kragen 4y agothe doomsday scenario is that the hashrate drop outpaces difficulty adjustment, because as you point out there's a rate limit on the difficulty adjustment; possibly in those 20 weeks the hashrate would drop another 90% because it isn't profitable to run the miners during those 20 weeks at the still high difficulty but much lower bitcoin price after two or three orders of magnitude drop in the hashrate 51% attacks would be quite affordable
- amluto 4y agoI would argue that they’re already quite affordable in a marginal sense — the marginal cost in power to supply 51% of the hash rate for long enough to attack the network is not terribly high. These attacks are not affordable in a capital sense when the miners are running near capacity. There isn’t an AWS with unlimited piles of hashing ASICS that you can rent for the hour at a reasonable price. You need to buy or lease these things, Ave capacity is finite. But I suspect that, if mining profitability per hash drops enough, then the capital limitation on attacks will go away.
- crote 4y agoThe problem is that the difficulty adjustment is done after a certain number of blocks, not a certain amount of time. If the hashrate halves that block count would take twice as long to reach, so you'd need to run at those higher difficulties for 40 weeks.
- kragen 4y agoalso a good point, but i think the scenario described would still be a big problem even without that exacerbating factor