6 ms·
It self-corrects slowly. The protocol adjusts the hash rate every 2160 blocks, or about once every two weeks, and the maximum negative slew rate is -75%. If su
by jiggawatts 4y ago
It self-corrects slowly.
The protocol adjusts the hash rate every 2160 blocks, or about once every two weeks, and the maximum negative slew rate is -75%. If sufficient miners pull out of the network within a fortnight or two, it's conceivable that Bitcoin will enter into a death spiral.
It goes something like this:
1. Miners realise that their operations are unprofitable, and do some mix of raising tx costs and/or turning off their kit.
2. The tx rate drops, so instead of one hash found per 10 minutes, it might be one every 20-30 minutes or even slower. The competition to get into these blocks heats up, further pushing up tx costs.
3. Bitcoin is no longer usable for small transactions, because the tx cost makes it a total loss. Small-to-medium users panic, as they realise that Bitcoin is not only plummeting, but that they can't get their money out without spending a significant fraction of it.
4. There's a rush to withdraw lump sums, creating a "run on the bank". This nearly instantly skyrockets tx costs.
5. Bitcoin value keeps plummeting, because nobody in their right mind would put USD into the system, and everyone wants to get out. Miners won't shut down their equipment at this point, because with high tx costs they're still profitable or nearly so. But they're paid in Bitcoin, so the real value of those tx fees is plummeting also.
If the above happens fast enough and with the right timing, with the most precipitous downturn happening right before a difficulty epoch, then the difficulty might not adjust enough downwards at the right time.
Then there's a chance that Bitcoin will go to 0 in the subsequent 14 days as miners walk away from their operations and users loose all faith that there's any value left in their coins.. and even if they do believe in that value, there's nothing they can do with their coins because hashes are found only every few hours and it costs the price of a new car to get into those blocks.
"It'll just adjust", you say. Sure... -75% after that fortnight[1], and if that's not enough, it'll be nearly a full month of an absolute crypto bloodbath. How long did it take Terra/Luna to crater? Days?
Caveat: Having said that, I suspect there's enough momentum in the system because of all the HODL-ers and self-interest in the miners. The miners have much of their earnings tied up in Bitcoin. If they let the system crash, they lose their profits. But there is a point where they'll realise they're throwing good (fiat) money after bad (crypto) and literally pull the plug and turn off the kit.
[1] Or longer! Remember, it's every 2160 blocks. If too many miners give up mining all at once, this could be over a month.. or longer.
- latchkey 4y agoETH adjusts difficulty every block. They also partially reward miners for 'uncle' blocks. IMHO, it was a good modification to the protocol. Theoretically, BTC could make a similar adjustment and avoid the doomsday scenario that you outline.
- baby 4y agoA good modification of the protocol is using a PoS as it just doesn't have these issues
- dboreham 4y agoBut it has other issues.
- latchkey 4y agoI'd love to see PoS be successful and end PoW mining, I just have my doubts that the current ETH devs can pull it off successfully. PoS opens up its own can of worms. Namely, it is vastly more complicated and not as well understood or executed in as large of a scale as PoW currently is. The Merge is honestly terrifying to me given the amount of money on the line. As a software engineer who's built large distributed systems. I've always relied heavily on having robust test suites so that I know that when I'm deploying code, there is little risk of failure. In this case, the devs have had years to get the test suite passing [1] and the plan is to launch in a few months. I really wish that this was looking a bit better than it is at this point. There also isn't a published set of plans for what happens if things go wrong. We could end up with a forked chain and ETC situation again and I'm not sure it would survive that. So yea... I'm optimistic, but I'm also a doubter. I'm more than fine with them taking as long as they need to get it right. The current rush seems entirely unnecessary to me. Especially since the devs claim they don't care about the price. [1] https://hivetests2.ethdevops.io/ https://hivetests2.ethdevops.io/
- cypress66 4y agoI think it will be delayed a bit if necessary but it is still clear that a working PoS is completely feasible. I had a lot of doubts too about the pace of dev but the shadowforks and ropsten testnet makes me thing it's going to happen. I think that even if the devs don't care about the price, they care about the image and they want to get rid of the stigma that eth is bad for the environment. When that stigma became so widespread is when eth went all in on the PoS development.