A Two-Round Proof of Work instead of PoW

[Epictetus]

I've got an idea, and I would like to share it with you here to discussed/ripped it apart:

Two-Round Proof Of Work:

We devised a process with two rounds we called Two-Round PoW: In the first round, we run a similar Bitcoin PoW with a small hash complexity for all the miners. And in the second round we run another Bitcoin PoW only among the winners of the first round.
Let's N1 be the number of all miners. The goal of the first round is to select the first N2 miners that succeed to resolve the first round PoW puzzle. In the first round, each miner will solve for its own ounce1. The second round competition is only among the winners of the first round, the N2 miners. i.e each one of the N2 miners will solve for its ounce2. The complexity of the second round can be made adjustable like in Bitcoin to maintain the block time.  

For example: if we have N1 = 100000 miners in the network, we can decide to chose to only work with N2=100 miners for the second round of TRPoW. We can set the complexity of the first round in such a way that it can be with minor electricity consumption. And then adjust the complexity of the second round. In addition, we can also adjust $N_2$ to maintain the block time. Like this we can reduce the electricity consumption by as much as we want. So here we have two rounds with two parametrable complexities. We can view Bitcoin PoW as a Tow-Round PoW where N2 = 1 and no complexity in the second round.

In short, In a Two-Round PoW system, all the miners compete in a first round where each miner solves for its ounce1 and the fastest N2 miners that solved the first round puzzle, i.e the winners of the first round advance to a second round where they get to choose between them the fastest miner to solve its ounce2.

For more details please read here. It is a very short algorithm: https://www.researchgate.net/publication/356188255_A_Two-Round_Proof_of_Work

or Here on google drive: https://drive.google.com/file/d/1XVwMuDHIEBYX2GbAI_rFzx8eEsM7Z5ps/view?usp=sharing

Anyway, I just wanted to throw the idea out here to see if there are any obvious reasons why it couldn't be implemented, and to hopefully spark a discussion amongst those better qualified than me.

Thank you very much in advance for your thoughts / suggestions / ideas / critics

[BlackHatCoiner]

This sounds like you're wasting more energy than you should. There are two rounds and there will be miners who'll solve the first PoW, but not the second and those who'll work for the second PoW, but they'll not solve it.

Could you give us the PDF in plain format and not via researchgate.net which requires us to sign up to read it?

However, even if I don't understand it that much, don't you consider it possible to have calculated something wrong? By that reasoning, we can have infinite rounds of PoW and no energy required.

[Epictetus]

Thank you for your quick reply. Once in the second round, the behavior is like Bitcoin PoW: If no one among the N2 miners finds a solution within the block time there will be adjustment to the complexity. etc. etc

I will attach the PDF. Here on google drive: https://drive.google.com/file/d/1XVwMuDHIEBYX2GbAI_rFzx8eEsM7Z5ps/view?usp=sharing

Thank you.

[garlonicon]

There was a coin with two phase Proof of Work called Luck and it failed, because it doesn't matter how many phases you have. Miners can always mine in a black box way. You can always implement mining as a single phase process and distribute rewards between N people by grouping them in mining pools. You know what happened in Luck, where people were trying to avoid pooled mining? People were disappointed, because mining a single block was taking more and more time, some miners received one block per month.

The problem is not splitting mining into more phases. The problem is to give each miner its own share. If someone solo mined N times easier block, that miner should receive N times lower block reward for that. The problem is: by doing that kind of things on-chain, you will flood the network with small transactions between miners. For now, centralized pools are handling rewards off-chain, miners just withdraw their coins later, when they accumulate some significant amount. So the whole problem in decentralized mining is about joining shares, joining transactions, and in general compressing all data related to payments for miners, so that could compete with centralized pools.

[Epictetus]

You can download my PDF directly here: https://drive.google.com/file/d/1XVwMuDHIEBYX2GbAI_rFzx8eEsM7Z5ps/view?usp=sharing

I think the Tow-Round PoW is different from the Luckcoin concept.

From what I read, it looks like Luckycoin was based on Proof of Work with minor modification in the complexity adjustment.
https://cryptomining-blog.com/tag/luckycoin-scrypt-crypto/
Or a mention here: https://www.goldmansachs.com/insights/pages/crypto-a-new-asset-class-f/report.pdf


Thank you for the input. Much appreciated.

[PrimeNumber7]

. We can set the complexity of the first round in such a way that it can be with minor electricity consumption.

TBH, you lost me here.

There is always the possibility that two blocks will be found by the same miner. How would you handle this?

Assuming you perfected the above issue, I would say that you should not reinvent the wheel. If however, you feel very strongly about your proposal, I would encourage you to create an altcoin with your proposal implemented. If it is superior to the status quo, the market will reflect your altcoin accordingly.

[Epictetus]

Yes. I totally agree with you. The best way to test the idea is to actually create a coin that is using the TRPoW.

Regarding the block fork. Here, we assume that we have a system like Bitcoin where we replace
PoW by TRPoW and everything else stays the same with minor adjustments due to the introduction of TRPoW.
Like in Bitcoin, a miner can decide to cheat and ignore the TRPoW instructions and the information that it receives from other nodes and continue solving for a block and take the reward for itself. It will create a fork. Later on, the miners in the network will simply go with the longest block and ignore the cheater block.

Thank you for your comments and the good suggestion.

[garlonicon]

From what I read, it looks like Luckycoin was based on Proof of Work with minor modification in the complexity adjustment.

Not Luckycoin, just Luck. Details are in this topic: https://bitcointalk.org/index.php?topic=5254068
Also there was some discussion about elimination of mining pools: https://bitcointalk.org/index.php?topic=5254327

I think learning something from their mistakes before starting a new project is a good idea.

[Epictetus]

Thank you very much for your reply. I read the thread. It is quite interesting indeed.
I was NOT able to download the Luck white paper because the link and the website are no longer working.

However, I read the different discussions here and It looks different from TRPoW.
Luck is using two phases of PoW to reduce the domination of large mining pools. While TRPoW is using two phases PoW but to reduce the network electricity consumption. Based on the few details in the discussions in the thread, It seems though that we are doing completely different things. TRPoW is not using any fancy formulae to calculate anything. Our changes are purely distributed algorithmic.   

Please share the white paper PDF here if you still have it.

Indeed it is important for me to learn from this experience. Also, I am not trying to reinvent the wheel as I was not aware of this project until you mentioned it.  

Thank you as always for your great sharing.

[odolvlobo]

I don't believe your proposal reduces energy consumption. Energy consumption in PoW is limited by the value of the block reward, and is not related to the difficulty of the solution. The difficulty maintains the time between blocks.

The goal of the first round is to select the first N2 miners that succeed to resolve the first round PoW puzzle.

How does the network decide which are the first N₂ solutions? Each miner could insert their solution to their list regardless of how many other solutions they have already received.

[Epictetus]

Hi Odolv,

It is explained actually in section 3 (Figure) and 4.

Anyway, In the TRPoW, each miner will hash the block header plus its public key plus an ounce repeatedly. The public key is added into the hash to avoid a miner using the result of other miner as its own result during the broadcasting of the result. The complexity of the first round is the same for all miners but each miner is solving for its own problem. In the second round, each winner miner will hash the block header plus its public key plus its ounce of the first round plus an ounce repeatedly.

Please have a second look.

The other byproduct benefit of using of the public key in the hash is to help reduce the domination of big pools. Though it is not the best solution as minors can still by pass it. The main goal of TRPoW is to reduce the electricity consumption.


Thank you for your comments and ideas

[odolvlobo]

You don't explain how your system reduces the energy consumption. Miners are going to expend as much energy as they can afford in order to be included in the second round. Right?

... and the fastest N₂ miners that solved the first round puzzle advance to a second round where they get to choose between them the fastest miner to solve its ounce2.

It may not be important, but how are the "fastest" miners chosen? In Bitcoin, the first block produced is not necessarily the one ultimately added to the chain. In your system, each miner can have a different set of first N₂ blocks. Also, just to complicate matters, as I stated before, there is no reason for a miner not to submit their block as one of the "first" N₂ blocks, even if they have already seen N₂ blocks.


BTW, the word is "nonce", and not "ounce".

[Epictetus]

The electricity will be reduced by the fact that Round 1 is not computing-intensive and most of the miners won't be "allowed" by the rules (if followed) to advance to the second round. The second round will be only among the first N2. This will reduce the electricity greatly.

In the traditional PoW, all miners are hashing for big complexities most of the time. While TRPoW round 1 complexity is not electricity intensive. The complexity of Round 1 will be set to be small. So Round 1 is pure Luck and Round 2 is where the calculation power is needed.

The other points that you mentioned, their treatment is very similar to how Bitcoin is dealing with these issues.

Keep in mind here that we are changing only the PoW. Everything else stays the same.

the blocks mined by cheaters will be ignored like In Bitcoin and the network will always work with the longest valid chain.

I agree with you that it is possible that we have two sperate groups with two different N2 miners that will produce two valid Blocks. But again the same situation is happening in BTC. So the same treatment applies here.  

thank your for sharing and the correcting of "nonce". Much appreciated.

[odolvlobo]

The electricity will be reduced by the fact that Round 1 is not computing-intensive

Regardless of the complexity, if I use more power to increase my hash rate, then I will have a better chance of getting to the second round. Right? So, why wouldn't I use as much power as I can to get into the second round?

I think a solution would be to eliminate the PoW in the second round and simply make it a random selection (somehow). That would cut the expected block reward by a factor of N₂, which would cut the power usage in the first round by that same factor.

[PrimeNumber7]

The OP appears to be proposing something that will result in only 0.1% of the miners using their equipment in the second round of mining, and the remaining 99.9% of miners will turn off their equipment during the second round.

The above is not possible due to pool mining. As an example:
If there were 10,000 pools participating in the 1st round, only 10 would make it to the second round. Once the second round starts, the miners who were mining for the 9,990 pools that are not participating in the second round would switch to one of the 10 remaining pools.

There are currently 4 major mining pools (with the likelihood of there being more that are not public). Most likely, they would all be able to participate in the second round of mining, however in the event that one pool is unable to participate in the second round, the miners mining for that pool will switch to one of the remaining pools. If a "solo miner" would be able to make it to the second round, it would make financial sense for them to "sell" the ability for others to "mine in their name" and if no pools are able to participate in the second round, all of the pool miners would "buy" the ability to mine in one of the solo miners who are participating.

tl;dr - total electric consumption will not be reduced by implementing the OP's proposal.

[garlonicon]

Please share the white paper PDF here if you still have it.

Wayback Machine still has it, as well as some lucknet.club pages: http://web.archive.org/web/20201201172513/https://lucknet.club/doc/luck.pdf

miner will hash the block header plus its public key plus an ounce repeatedly. The public key is added into the hash to avoid a miner using the result of other miner as its own result during the broadcasting

But you know that Bitcoin blocks are produced in a similar way? Small miners include public keys for mining pools, so even if all blocks will be signed by miner's keys, it won't change anything. You can have one phase, two phases or N phases, it doesn't matter. Pools are going to download your client and behave like any other network participants. Then, they will be connected with hundreds of nodes and rewarding them "per share", no matter how complex that single share will be and how many phases will be needed to get it.

The electricity will be reduced by the fact that Round 1 is not computing-intensive

If you have something that is not computing-intensive, then the energy consumption is reduced, but along with security. I guess the same thing could happen as in Luck: if you have two phases and they are not equally hard, then miners can game the system by dividing it differently between those two phases than you did in the original software. For example, your official client can mine the first phase for a short time and spend the most of the time on the second phase. But some optimized version could do the opposite: spend a lot of time on the first phase to pretend there are millions of miners, and then quickly solve the easiest blocks by completing the second phase.

Or maybe they will find an equillibrium somewhere in between, maybe with the same amount of time for each phase, maybe 2/3 for the first and 1/3 for the second one, it really depends on your network parameters. In Luck it reached the highest limits they chose, as far as I remember their "alpha" value exponentially increased to the maximum value and stayed there, because that case was the most profitable for huge miners. In their first version there was no such limit, so without their first hard-fork, the network would be dominated by the biggest miners forever.

most of the miners won't be "allowed" by the rules (if followed) to advance to the second round

You cannot assume that all miners will follow the rules, just because you placed them in your client. Rules have to be wired in the protocol to be followed, if some rule is only client-specific, it can be bypassed. For every successful coin you can find custom miner software, where miners optimize things far beyond what you gave them in the official version.

Also, if there are millions of miners and you randomly select 100 of them, you encourage the rest of the miners to form mining pools. Then, you will end up in a situation where 100 miners are selected as before, but there are only 100 huge mining pools, they always get everything and distribute that to smaller miners. Almost no solo miner will follow your system if that miner will be rewarded once per month, because the risk of price drop after entering exchange is higher than the potential reward from solo mining.

[larry_vw_1955]

tl;dr - total electric consumption will not be reduced by implementing the OP's proposal.

I'm sure the miners wouldn't be supporting such a thing because it just makes it even harder for them to get a reward.

Also, if there are millions of miners and you randomly select 100 of them, you encourage the rest of the miners to form mining pools. Then, you will end up in a situation where 100 miners are selected as before, but there are only 100 huge mining pools, they always get everything and distribute that to smaller miners. Almost no solo miner will follow your system if that miner will be rewarded once per month, because the risk of price drop after entering exchange is higher than the potential reward from solo mining.

I wonder if the people that wrote that paper realized that. that's thinking ahead! the bitcoin mining system is really hard to make changes to without causing things to get "out of whack"

[Epictetus]

The electricity will be reduced by the fact that Round 1 is not computing-intensive

Regardless of the complexity, if I use more power to increase my hash rate, then I will have a better chance of getting to the second round. Right? So, why wouldn't I use as much power as I can to get into the second round?

I think a solution would be to eliminate the PoW in the second round and simply make it a random selection (somehow). That would cut the expected block reward by a factor of N₂, which would cut the power usage in the first round by that same factor.

If you use more CPU power, you have always better chance to advance to Round 2. However, because Round 1 is not CPU computing, small players will be able to advance to Round 2 too. It is like we are giving more chances to small players to earn.

About your second point, let's me reflect on it.

[odolvlobo]

If you use more CPU power, you have always better chance to advance to Round 2.

So then, why would miners use less power?

You state "In the first round, we run a similar Bitcoin PoW..." If the first round is similar to Bitcoin PoW, then it has the same economic incentives, in which the cost of mining is limited only by the value of the block reward, regardless of the complexity. It seems to me (I may be wrong) that you may be succumbing to a fallacy similar to the belief that more efficient miners (in terms of J/hash) in the future will cause the overall power consumption of mining to fall.

[Epictetus]

In an ideal world with TRPoW, most miners will spend their time moving from one Block Round 1 to another Block Round 1 because the round 2 is limited only to the fastest few N2 miners. And hence we will able to reduce the electricity from not involving everyone to go on a spree of CPU intensive hashing all the time.

So the idea is not to advice for LESS CPU power but to create a system that do the same thing like BTC with less electricity consumption.