Hacker News Reader

Comment by konschubert | Hacker News Reader

The article is misinformed:

I am the first one to jump on Bitcoin's flaws, but this here just reeks of somebody not understanding the protocol:

> As I argued, once Bitcoin’s price falls below its cost of mining, the incentive to mine will deteriorate, thrusting bitcoin into a death spiral.

The cost of mining adjusts depending on the hash rate, so if nobody is mining the cost of mining goes down until people start mining again.

EDIT: The mining difficulty gets adjusted only every 2013 blocks, so it might be that mining activity drops so low that the block chain starves out before the difficulty can be adjusted down.

Is this what the article is trying to say?

View on HN

The article addresses that further down. frankly I don’t quite know if their argument is valid. I am skeptical, as bitcoin existed before it got any significant value.

“. Similarly, when the price of bitcoin falls and miners exit, the cost of mining decreases. However, the number of miners cannot fall below a certain level, because without the miners providing the computing power to maintain the ledger, the bitcoin blockchain will not remain viable.

Mining at a cost higher than the cost at which you can sell in the futures market destroys value. So, any rational investor — even one who strongly believes the price of bitcoin will rebound — has no incentive to mine if the cost of mining is higher than the future price and is better off buying in the futures market. ”

But buying only makes sense if you can "realize" the gains - for that, the BTC network (i.e. transfers between accounts) has to continue working. In the absence of mining gains from newly-minted coins, rich individuals that want to use their wealth will be paying for mining in transaction fees.

I suppose one source of risk is that if enough miners exit then it becomes viable for a bad actor to purchase enough computing power to mount an attack.

IIRC 51% attack happened on one of the bitcoin forks.

The cost of mining only adjusts after a set number of blocks. It's possible for there to be a sudden drop in price, meaning the majority of miners shut down unprofitable rigs, and then it takes essentially infinite time to mine the required N blocks to adjust the difficulty.

Wouldn't it pay for those then sitting on hardware to mine at a loss, in order to adjust the difficulty, in order to restore the value somewhat.

I'd guess it would get very spiky as people tried to buy up preempting the "rescue" efforts.

This has been addressed in the article:

> So, it appears bitcoin is now entering a death spiral: If the price continues to drop and the cost of mining does not fall correspondingly (the cost of mining will algorithmically decrease, but not necessarily to same extent as the decline in prices), bitcoin will quickly go to zero.

Why would the cost of mining not fall correspondingly? It will adjust until it becomes profitable again, nobody is going to run mining farms if they are clearly losing money every single day.

The way the bitcoin protocol is designed, the mining difficulty and rewards only adjust after N blocks have been mined. It doesn't respond in real time to the hash rate on the network.

Because the adjustment window is not real-time-based, but block-based: every 2016 blocks, an adjustment is done - with the obvious caveat being that the duration it takes to mine 2016 blocks is based on the global hash rate. Also, the downward adjustment is limited to (IIRC) halving the difficulty at max.

So, given the number of miners drops too far too fast, Bitcoin may end up in a situation in which the global mining power is too low to even reach the end of the adjustment window in a realistic timeframe, which in turn triggers even more miners to shut off their equipment and sell their coins, which lowers the price even more and lengthens the timeframe, which in turn...and so on. This situation is known in Bitcoin circles as the "chain death spiral", and it has been a purely theoretical thing - until now, at least.

> Because the adjustment window is not real-time-based, but block-based: every 2016 blocks, an adjustment is done ...

This is definitely true. Luckily for miners there are two other blockchains their mining equipment can make use of which don't suffer from the slow difficulty adjustment. They are both version of Bitcoin Cash (BCH and BSV) and they both have 10-block adjustment periods. This makes them much more resilient to sustained periods of depreciation.

You're implicitly assuming that bitcoin will retain value. If the value of bitcoin halves every 6 months and the cost of mining halves every 7 months, bitcoin will never* be worth mining again.

* Of course it wouldn't be a smooth decline so there may be short periods where it occasionally becomes viable.

Where is the author pulling their mining costs from? I'd love to see their source. Mining costs are mostly concentrated in purchasing the equipment and setting up a factory. Once that's done, your day to day expenses are electricity and labor.

Many of the biggest Bitcoin mining operations are in places like China, near hydro-electric dams, where the electricity cost is extremely low. Labor is also cheap. Since they've already paid the setup costs, they are not going to be shutting down even at $1000 BTC.

There have already been hash rate wars. BCH has tried sending BTC into a death spiral before, by luring over miners to their blockchain. It worked, for a bit. Eventually, transaction fees went up on BTC, and miners were incentivized to return to BTC.

The average block time is 10 minutes for Bitcoin. Hash rate adjusts every 2016 blocks. That's 14 days on average. If the hash rate dropped 50%, it's 28 days.

Miners probably have a bit of a sunk cost fallacy, since there's a fixed cost associated with acquiring the hardware. Plus, if they're still bullish on bitcoin, the short term loss is probably worth it in their eyes.

I think you nailed it, although I don't think miners even need to fall for a fallacy for your argument to work: some major miners have made big capital expenditures (efficient hardware, cooling, etc.) to keep variable costs low over the long term. Even if the price drop means they'll never mine themselves out of debt, it could still be the rational thing to do for them to keep the machines on, as long as their reduced variable cost is still below than the market price.

Aaah. So he's saying that the price might fall so quickly and the hash rate might drop so quickly that the difficulty cannot adjust because they blocks to adjust the difficulty won't get mined?

(The difficulty gets adjusted only every 2013 blocks)

Which won't happen since the 'cost' of mining is electricity and hardware costs which for some miners is probably very close to zero. All this is going to do is shuffle out the miners who have no efficiency.

Sounds like an opportunity for someone to get control of more than 50% of mining capacity.

At which point, could you unilaterally hard-fork it, e.g. to fix this problem with the difficulty-adjusting rules?

Though if that is an option, I would guess it would be more likely for a coalition of miners to choose to do so before any individual achieves 50%.

Not only this, but the article also doesn't even mention transaction fees to rule out that they may offset a lacking mining reward.

this article is misinformed. there is no single mining profitability threshold and some miners will mine at a loss. some miners are profitable at a very low price.

Exactly. I don't understand how the article can miss such a fundamental aspect of Bitcoin's design. What this means is that the cost of mining will go down and adjust accordingly.

Reply on news.ycombinator.com