Creating artificial delays for fairer mining among nodes with high latency

For miners, it's not just about having affordable electricity and powerful hardware, you also need solid connectivity with low latency to communicate with other nodes. If your blocks are consistently late, they're just gonna get rejected. I came across this article about how network delays affect Bitcoin mining, and the authors concluded that latency is crucial. It seems to me like this could be tackled with some technical solutions, and I'd love to explore this further with you all. Here are some ideas I've got so far: 1. Instead of instantly using the first solved block, we could wait for a few other blocks to come in during a certain timeframe. Then maybe we can either reward the first block while also giving some rewards to the others. This way helps spread out new coins and gives miners some income for their efforts. Ethereum kinda does this, right? But with late blocks getting lower rewards, which feels unfair for those mining from far away. They don't get rejected, though; they’re tagged as "uncle blocks" and combined later. No wasted hash power there, which is a plus. 2. Another option is to adopt a Directed Acyclic Graph (DAG) approach like Kaspa, where all mined blocks are accepted as valid. I just stumbled upon this currency while doing some research, so I haven’t dug deep into it yet. But it sounds pretty cool that you won't waste electricity, and your reward stays the same regardless of your latency. 3. We could also just pick a random block as the winner, giving even the last one a shot at winning. Sure, adding this artificial latency could slow things down a bit. But on the flip side, it could make the system less reliant on pools, spread miners more geographically (including regions with lower costs and renewable energy), and make it more accessible (even slower ASICs could work). Implementing this could be tricky for Bitcoin, but who knows, it might be feasible in other systems.

Do you mean this article/paper, https://www.researchgate.net/publication/353719275_Characterizing_the_Impact_of_Network_Delay_on_Bitcoin_Mining? I only skimmed your idea, but i immediately notice this part is flawed. Miners would just use their ASIC to mine alt-coin, which means total electricity used by miners remains same. WhatToMine shows there are 5 SHA-256 based altcoin that currently is more profitable than Bitcoin.

Yes, that article. I didn't linked it because for me it was interesting only it's conclusion. A good catch about mining of alts, didn't thought about it. Generally if the idea can be useful then other currencies can also use delays so the overall electricity consumption may be decreased. As a separate idea, what if give some coins even for undermined blocks, i.e. if you found a hash with only 30 zeros instead of 31 required by the current difficulty level. You still spend an electricity so you are not a fraud so you can get some reward. The problem that I see is that in such schemas it would be problematic to find and transfer all the blocks, but with a delay this is possible. In the Bitcoin Unlimited there was a similar idea to also use intermediate blocks for weak confirmation https://bitcoincashresearch.org/t/spearheading-the-future-implementation-of-weak-blocks-intermediate-blocks-for-bch-proof-of-work/1333 Generally speaking with the delay we can adjust speed of blockchain without difficulty adjustment. This may be a useful thing. And speaking about the proof of location I wish to add that theoretically we can also use a time from satellites. Then you are not affected by connection problems to a local NTP server. The GPS provides time it but it's not signed. A more weird idea is to use signals from stars that are seen in different parts of globe but that's not really useful.

Have you done research about uncle/ommer block on ETH when it used PoW? Some people critic for some reason, such as 1. It makes ETH supply/production rate higher than expected. 2. It somewhat burden ETH node, since more data is being transmitted and uncle/ommer block header need to be stored on blockchain. I don't see how it's possible when there's no way to know and verify actual hashrate capability. And IIRC average time to mine a block is determined by either poisson or exponential distribution. That would introduce some centralization, where i expect such idea wouldn't be welcomed by Bitcoin community.

Thank you, are absolutely right. When I did research I found that the Uncle Block actually even was abused for years by one of the biggest pool. See the video "Aviv Yaish (HUJI) - Uncle Maker: (Time)Stamping Out The Competition in Ethereum". And there was mentioned that also other smaller currency was attacked. This only added me more prejudices against pools. This guys should be interested in support of system but instead abused it. The uncle blocks intended to at least slightly reduce a dependency on pools was abused by them. See, if we have a delay then nodes will have a time to wait for all blocks and select one without creation of uncles. But then transactions from the late block needs to be repacked to avoid their loss. Not sure if this possible, maybe that was a main reason of uncles. Anyway, with a delay nodes have a time to exchange more data and check for double spending. With a longer delay they even can make a DHT lookup by an address to find its current transactions. Right now users have to wait for a few confirmations because their transaction may end up into a side block that later will be rejected. With a delay this can be reduced because we'll always know that only the one block is currently exists. Yesterday I found that the Nano uses the PoW more like a protection from Sybil Attack but instead have more node that are expected to be mostly honest. It's hard to find incentives for miners, but it looks like the system works. Still it may be relatively easy to attack it, but transactions log allows to recover. The Nano uses DAG and this looks like a good fit. I'll learn more on the currency once get more time. To recap, in the Bitcoin the PoW plays a few roles: 1. To determine the first winner by producing the longest chain. With a delay we can do this more intelligent i.e. use different strategies, make all winners, use DAG or uncles etc. 2. To make a price bigger by electricity consumption. Questionable, TBH. Maybe if miners had fewer expenses they wouldn't sell their coins to pay bills making inflation smaller. 3. To regulate speed of coin producing. Again, delays can handle this more precisely. 3. As a protection from attacks like spam, 51%, Sybil etc. It looks like even with a smaller hash rate this attacks never affected Bitcoin and only a few other currencies were attacked (at least from what I remember). So maybe the PoW difficulty can be significantly reduced. With a big delay it should be possible to make user wallets to become a miners themselves. Your phone can receive a block to verify and perform a short PoW. If this happens only a few times per day users won't even notice. The ASIC miners will never become profitable. This is entirely new quality that wasn't possible before. Like a PoS without a stake. More like a voluntary BitTorrent network. Another question if users will need for a such slow system. The answer is probably no, maybe only for things like online cinema subscriptions. But this is an example that the delays may be useful as tool in building distributed systems. We may change properties to achieve qualities that we want. Unless we have an extreme protection from censorship we may relay on external systems (DNS, TLS Root Certs, NTP, etc.). This risk is not that big, and BTW, if compare, definitely smaller than risk of pools that erode decentralization.