I think that implementing the protocol described in this paper wouldn't do much to diminish the total energy spent mining, but it would increase network instability and the potential for malicious miners to game the network to their advantage.
This protocol creates a condition where every other block actually requires that the miner to find 2 valid blocks in a row (the election block, and then the second round block), with a fall-back timeout of x minutes coming into effect after the first election block is found. Before x minutes has passed, the election winning miner(s) will have much less competition in trying to find the new block, so it seems to me that it is in their best interest to not publish their election block, until they also find a second round block! This way, the timeout condition will never be hit, but this also means that all other miners on the network will be working to mine their own election-blocks (and second round blocks), with little incentive to stop the search.
With this in mind, it is unclear if any rational miner would turn off their mining hardware ('mining-save-mode'), especially when considering that a miner will have many costs, including overhead (rent), staff, electricity, etc. Turning off their mining hardware will mean completely forgoing their chance of any revenue (the block reward), while only dropping their electrical consumption (and not any of their other costs). This only serves to put a high opportunity cost on any time spent not-mining.
Other issues I can see:
- the bitcoin network has no mechanism for registering/identifying individual miners over time, so it is unclear how the protocol could restrict the eligibility of miners participating in the second round in each mining epoch.
- The paper claims that the miner which wins the first round of the epoch would not
be allowed to participate in the second round, but it is unclear how
such a thing would be enforced. What is to stop the first-round winner from also attempting to find a 'runner-up' block and compete in the second round? Even if miners were somehow made to register an identity with the network (how such a thing would be possible is a big question in itself), what is to stop a miner from registering many identities, and switching between them?
- There is no way to reliably enforce the time-out process for the second round. The network is asynchronous, it is not possible to accurately define '10 minutes after the election-block was found' as a starting point for the fallback process of allowing all miners to participate in the round to help guarantee network liveness. This creates instability at the chain-tip, as the validity of blocks created by the fallback process may be unclear to network participants.
- This protocol may create incentives for miners to coordinate externally. For example: if I am a miner and I find a valid election-block, then I may want to sell the ability to participate in the second-round block to other miners! Doing so could help normalize my expected returns, while allowing other miners to reduce their opportunity cost of otherwise not mining in the second round.
- Supposing the miners do not perform the election-block-withholding attack described above, in the event a single miner finds an election block, they may be further incentivized to perform a block withholding attack using their round 2 block as well! With little competition for the round 2 block, the miner would enjoy a much higher chance of having the withholding attack succeed, and they could begin searching for the round-1 block of the next epoch while withholding the round-2 block of the current epoch.
These are just some of my initial thoughts, I'm open to criticism of them, and there may be further issues I haven't seen. But at this time, I'm seeing a lot of problems that make this proposal somewhat uninteresting and untenable.
Perhaps also worth mentioning, I think that the premise of the paper is erroneous in the first place. The energy spent by bitcoin miners provides a mechanism by which the network is secured. Valuable things require security, and security cannot be gained for free. The openness of the bitcoin network simply allows a rather objective quantification of the energy (ie the opportunity cost) used to secure it, compared to other systems which have a more obfuscated or social form of security, that comes at a higher relative opportunity cost!
