One benefit of ASICs (and any specialized mining hardware) is that they decrease the threat of this problem. The marginal revenue for an ASIC miner is far higher than the marginal cost--most of the cost is upfront. ASIC miners won't turn their hardware off even if their profit falls precipitously.
On the other hand, anybody mining with a CPU, with marginal revenue close to marginal cost, will shut off the program as soon as revenue falls. This increases susceptibility to this failure as well as other attacks, especially by a 'terrorist' who begins mining at a loss in order to force all of the other miners into negative profit so he can monopolize the network.
On the other hand, anybody mining with a CPU, with marginal revenue close to marginal cost, will shut off the program as soon as revenue falls. This increases susceptibility to this failure as well as other attacks, especially by a 'terrorist' who begins mining at a loss in order to force all of the other miners into negative profit so he can monopolize the network.
I disagree with you. Forget about CPU/GPU mining. Marginal revenue will be very small for ASIC miners in future due to competition. But cost of mining will be very high since Bitcoin will be very expensive. Thus even small decline of bitcoin price could incur significant losses to miners and paralyze the network. Most miners won't be able to tolerate significant decline of price. Please see again my calculations in the original post.
Words like marginal revenue make no sense in that context.
LIFETIME Margins are likely to be very low but lifetime margins include CAPITAL COSTS + OPERATING COSTS. Operating margin (mined value vs operating cost) is likely to be very high otherwise hardware will never be deployed to begin with. The operating margin has to accommodate the amortized lifetime hardware costs. The Capital cost is a sunk cost, it has already been paid for. The only way to get a return on it is to keep mining. Now operating margins might be squezed from time to time but if you stop mining then you turn that sunk cost into a realized loss.
Still even if a miner wanted to quit the probability that all miners will all have the exact same operating costs and simultaneously all stop mining at the exact same time is essentially zero. If that happen then Bitcoin will likely be hard forked with a different difficulty adjustment algorithm.
You also need to consider that some miners will simply be unable to stop. The lowest priced power is available to industrial users and in massive amounts (think >200 KW continual load). The power grid can't react to massive changes in these large loads so most industrial power contracts have penalty clauses. If a massive mining farm uses 200 KW (at the lowest possible energy rates) and then shut it down THEY STILL PAY. Their monthly bill for the next 3 to 6 months is HIGHER at 0 KW then at 200 KW. Simple version is that shutting down overnight = even higher costs then operating and revenue goes to zero. There is no scenario which would result in a greater loss for the miner.
