PoW heavy computation is a feature, not a bug --> please critique

[juvo_3]

Hello.  I am new.
I have been thinking about the issue of Bitcoin's seemingly indirect relation between BTC price to computational difficulty and therefore electricity consumption. The other day I came to this realization and I would like people to critique it:

One of the issues that caused the correction in price this year was that bitcoin technically lost utility to be used in small transactions. This was because the fee to effectively place information on the ledger was quite high due to high request volumes. Another issue was that the algorithm difficulty of the bitcoin protocol was so high due to the proportional hashing power available to the network. This caused the electricity consumption of the network to exceed that of a small country. Surely there are other factors for the price correction, but let’s just keep to those two for now.

People have said the PoW algorithm is the cause of this, and to an extent it is true. A miner has the incentive to mine based on profit. When they mine a block, they obtain the transaction fees and the reward for that block. With this profit, they pay their overhead and manage the net profit as they wish. At this point in time, the reward of the block is 12.5 coins. Correct me if i am wrong, but there is about 2000 transactions in a block, each of which will be associated to the highest fee transactions in the request pool.

Therefore:
block_profit BTC = block_reward BTC+ trans_feesBTC

The issue that we faced, at least from what I could tell, was that when a miner created a block and the price was 20,000USD per bitcoin, that’s how much their gross income was scaling ((12.5+trans_fees)*20,000USD). The incentive of the miner is to be competitive to other miners (which is the doing of PoW), so the game theoretic choice would be to increase the hashing power they contribute to the system with the profit they made. As a result, the hashing power went up and the bitcoin protocol therefore increased difficulty which lead to more electricity consumption. My initial thought was that due to the competitive, game theoretic behavior of participants when using the PoW algorithm, Bitcoin's price would have a barrier based on electricity consumption. If the price increases, the miners will just increase the mining capacity on the network because they are competitive as per the PoW algorithm.

I find now, this is wrong, and I am under the impression that other people maybe following the same logic (OR... Ia m a simpleton and just made a realization that everyone already knew)
What I have come to understand:

The bitcoin protocol is still in the phase of distributing coins into circulation. When mining first started, we got 50 coins as a reward. At this time the transaction fees were minimal as the use of Bitcoin blockchain real-estate was not so high in demand. The protocol states that the reward will halve after every 200,000 blocks or so. Eventually the reward value will decrease to an infinitesimal amount as we approach the maximum amount of Bitcoin that will be in circulation. This means that the miner's profit from the block will dominantly be in the form of transaction fees.

The reason why we had excessive electricity consumption of the bitcoin network was because the demand of Bitcoin was excessively high at the time. And the total profit of the block consists of the block reward of 12.5BTC + trans_fees. Just the gross income from the block reward was 12.5*20,000USD =250,000USD every 10 minutes. The PoW algorithm in a way regulated this price increase with high computational power demand. It actually seems like a feature and not a bug.

As the block reward depletes, the difficulty of the network with be regulated in a game theoretic manner. Here, the game players are the end users (people making transaction requests). The game they will play will be how much fee should they put down and the outcome of the game is whether their transaction will be selected from the transaction pool by miners. As the volume of transaction requests increase, I would expect to see the players put more down in fees. And of course, the end users will be observing the fees placed by their competitors. The profit made by the miner from these fees is what will dictate the computational power delivered by the miner network. (at this point, the word miner will also be somewhat diminished).  

This game between end users will also be effected by secondary layers such as the lightning node network. Lightning nodes will be placing transactions on the Bitcoin blockchain that summate a multitude of transactions. So the overhead cost of a lighting network node would include the payment of fees to the Blockchain miner network.

In this scenario, the electricity consumption and network difficulty seems to be a function of transaction request volume and the game theoretic transaction fee amount. Which is doesn’t seem to be a bad thing.  But I would like to hear others thoughts on this. It is the block_reward that is skewing the miner’s behavior to place more hashing power on the network which made PoW seem unfit. But at the moment, to me this seems like the protocol penalizing excessive demand and price whilst in the token distribution phase.

[juvo_3]

Okay, maybe my post is too long. I just want some critique if my thoughts are correct and if I have missed some factor.

in a quick summation of the above post. people have been criticizing the PoW algorithm because it is computationally heavy and therefore cannot scale in price or volume.

from what I have observed, the miners increase hashing power to the network based on the amount of profit they make. This profit is the block reward and the transaction fees accumulated.

as the block reward reduces over time, the profit per block will be the transaction fees. So the transaction fees will regulate how much electricity is consumed by the miner network. Therefore PoW actually has good regulation when it comes to price determination.

is that correct? is there any other factors that play into a miner's choices of mining?

the only scaling issue that I see is the transaction per/sec which is being currently addressed with the Lightning Node Network.

I would like to start this conversation because there are more questions to ask.
i.e. what happens if the transaction fees are low and the miners don't have the incentive to place hashing power on the network that secures it?

[amishmanish]

Hello.  I am new.
I have been thinking about the issue of Bitcoin's seemingly indirect relation between BTC price to computational difficulty and therefore electricity consumption. The other day I came to this realization and I would like people to critique it:
...

The reason why we had excessive electricity consumption of the bitcoin network was because the demand of Bitcoin was excessively high at the time.

Electricity consumption doesn't really have any significant relation with number of transactions or demand. Bitcoin miners are ASIC based devices with fixed power consumption. For example, the Antminer S9 is rated at about 1300W.

The PoW algorithm in a way regulated this price increase with high computational power demand. It actually seems like a feature and not a bug.

Umm. This is neither a bug nor a feature. Its more like a limitation at present because of the trade-offs that have to be made between block size and decentralization.

It is the block_reward that is skewing the miner’s behavior to place more hashing power on the network which made PoW seem unfit. But at the moment, to me this seems like the protocol penalizing excessive demand and price whilst in the token distribution phase.

PoW doesn't seem unfit in any way if you consider the following:

a.) Any form of alternate system like PoS or DPoS means the people with highest stake in the network can control it. Takes us back to the failings of our representative democracies.
b.) All the alternate systems claiming to be decetralized (Like IOTA ledger, Raiblocks etc.) make claims they cannot fulfill and are regularly releasing scams rather than features


I think you are trying to link Electricity consumption as you heard about the news articles about bitcoin's enormous electricity consumption. Well, Those figures are debatable and far from accurate. There was an upsurge in "Bitcoin's Environment impact" posts during that time. Here's a link where I tried to reason that the figures were inaccurate. Even if you assume that electricity consumption increased because more miners started mining bitcoin due to higher reward, that has no effect on Bitcoin price.

If you are new then a reading of Mastering bitcoin by A. Antanopoulous is recommended. There is a lot of FUD out there. You can look at some of my post history to see the kind of scenarios people come up. Happy Learning!

[juvo_3]

Thanks for the reply!

Electricity consumption doesn't really have any significant relation with number of transactions or demand. Bitcoin miners are ASIC based devices with fixed power consumption. For example, the Antminer S9 is rated at about 1300W.

The PoW algorithm in a way regulated this price increase with high computational power demand. It actually seems like a feature and not a bug.

Umm. This is neither a bug nor a feature. Its more like a limitation at present because of the trade-offs that have to be made between block size and decentralization.

So does this mean that no matter what the hash rate of the ASIC device, it uses the same amount of power? My relation of hashing power to bitcoin price is that the miner has more profit to either invest in more ASIC machines or (what now maybe a false reason, given your statement) run machines at a higher load and consume more electricity to increase hashing rate.

My understanding is that the miners have a control parameter which is "how much hashing power do I provide to the system" and that choice of control is governed by the profit of the block (profit = gross profit from block creation - [electricity + other overheads]). The gross profit from block creation is made of two functions:
- constant block reward (currently 12.5 and will halve in the future untill it becomes very small)
- transaction fees (highest fees picked from the mempool)

from this reasoning, there is currently a significant relation between electricity consumption and price of BTC, given that miners have this control parameter. But my thesis is that when the block reward diminishes, then the relation will become related to transaction fees (which should be governed by the number of transactions.)

So i guess what i want to talk about is whether what we perceive as a "limitation at present" will be the same when the block reward is no longer there (To clarify, not the limitation of transaction per second).

It is the block_reward that is skewing the miner’s behavior to place more hashing power on the network which made PoW seem unfit. But at the moment, to me this seems like the protocol penalizing excessive demand and price whilst in the token distribution phase.

PoW doesn't seem unfit in any way if you consider the following:

a.) Any form of alternate system like PoS or DPoS means the people with highest stake in the network can control it. Takes us back to the failings of our representative democracies.
b.) All the alternate systems claiming to be decetralized (Like IOTA ledger, Raiblocks etc.) make claims they cannot fulfill and are regularly releasing scams rather than features

I don't know enough of these other systems to agree or disagree with you on this. Id rather take incremental steps in learning rather than jumping the gun to say other solutions are scams or not. Although, my current understanding of PoS does lean towards the idea that you get rewarded for hording tokens from circulation which in of itself is not providing a utility to the network. This, off the cuff seems bad, but i am still getting my head around it.

I think you are trying to link Electricity consumption as you heard about the news articles about bitcoin's enormous electricity consumption. Well, Those figures are debatable and far from accurate. There was an upsurge in "Bitcoin's Environment impact" posts during that time. Here's a link where I tried to reason that the figures were inaccurate. Even if you assume that electricity consumption increased because more miners started mining bitcoin due to higher reward, that has no effect on Bitcoin price.

Yes, my first point of reference was the news articles. But surely there is a scaling of electricity consumption with hashing power in the network? I don't want to claim that the values in the articles were accurate. But isn't there more electricity consumption? and how does it scale with PoW? And how will it scale in the future?

In regards to my assumption, I am not assuming that the consumption of more electricity causes a change in Bitcoin price. I am suggesting that the Bitcoin Price will effect the miners profit which will change their game theoretic choice to put more hashing power on the network. So higher BTC price results in Higher Hash Rate when the block profit function is dominated by the constant term of block reward. I hope I am making sense.

If you are new then a reading of Mastering bitcoin by A. Antanopoulous is recommended. There is a lot of FUD out there. You can look at some of my post history to see the kind of scenarios people come up. Happy Learning!

Its a big book, I am slowly going through it.

[ranochigo]

My understanding is that the miners have a control parameter which is "how much hashing power do I provide to the system" and that choice of control is governed by the profit of the block (profit = gross profit from block creation - [electricity + other overheads]). The gross profit from block creation is made of two functions:
- constant block reward (currently 12.5 and will halve in the future untill it becomes very small)
- transaction fees (highest fees picked from the mempool)

from this reasoning, there is currently a significant relation between electricity consumption and price of BTC, given that miners have this control parameter. But my thesis is that when the block reward diminishes, then the relation will become related to transaction fees (which should be governed by the number of transactions.)

So i guess what i want to talk about is whether what we perceive as a "limitation at present" will be the same when the block reward is no longer there (To clarify, not the limitation of transaction per second).

Miners can definitely control the amount of hash power that they want to contribute, most likely by turning on and off the miners that they own. I'm pretty sure they can try to control the clock speed that their miners run at. The problem is that miners don't really mess with them. ASICs do have a specific frequency to run at such that they have the maximum efficiency. Each miners do cost miner to keep and they should be on at all times. Else, it is a net loss and a dead weight to the miner.

Yes, my first point of reference was the news articles. But surely there is a scaling of electricity consumption with hashing power in the network? I don't want to claim that the values in the articles were accurate. But isn't there more electricity consumption? and how does it scale with PoW? And how will it scale in the future?

Yes. While more hashrate does mean that the electrical consumption will increase, it is assuming that the hardware doesn't change.

In regards to my assumption, I am not assuming that the consumption of more electricity causes a change in Bitcoin price. I am suggesting that the Bitcoin Price will effect the miners profit which will change their game theoretic choice to put more hashing power on the network. So higher BTC price results in Higher Hash Rate when the block profit function is dominated by the constant term of block reward. I hope I am making sense.

That's somewhat correct. The prices is definitely partially influenced by the miner since they do sell some Bitcoins. However, more often than not, the prices is not representative of how much hashrate there is on the network. There is no metric to compare against.

[juvo_3]

Thanks for those answers!

Miners can definitely control the amount of hash power that they want to contribute, most likely by turning on and off the miners that they own. I'm pretty sure they can try to control the clock speed that their miners run at. The problem is that miners don't really mess with them. ASICs do have a specific frequency to run at such that they have the maximum efficiency. Each miners do cost miner to keep and they should be on at all times. Else, it is a net loss and a dead weight to the miner.

With this in mind, the only time a miner would have the incentive to shutdown a ASIC miner would be when the profit of the block generation would not even cover the electricity consumption in the case they won it? otherwise it would be kept on at a static hash rate per ASIC machine with a fixed power consumption. Awesome information. thanks.

Yes. While more hashrate does mean that the electrical consumption will increase, it is assuming that the hardware doesn't change.

What do you think about the hash rate supply once the block reward becomes small and transaction fees dominant? I don't see an incentive for end users (tx requester) to place a higher fee if the transaction volume was to drop (e.g. when transactions are executed on the LNN leaving less volume). At this point the miner would need to feasibly supply hashing to the network, but there is no feedback for end-users to increase fees to maintain the security of the network.

or have i missed something?

In regards to my assumption, I am not assuming that the consumption of more electricity causes a change in Bitcoin price. I am suggesting that the Bitcoin Price will effect the miners profit which will change their game theoretic choice to put more hashing power on the network. So higher BTC price results in Higher Hash Rate when the block profit function is dominated by the constant term of block reward. I hope I am making sense.

That's somewhat correct. The prices is definitely partially influenced by the miner since they do sell some Bitcoins. However, more often than not, the prices is not representative of how much hashrate there is on the network. There is no metric to compare against.

I agree, there is some influence that the miner has due to the process of them liquidating coins to run their operation, but the market price determination involves way more factors. What I am more concerned with is the behavior of the network once the coin supply starts to plateau and we have lightened the load of transaction volumes from the blockchain.

[Random Seller]

Thanks for those answers!

With this in mind, the only time a miner would have the incentive to shutdown a ASIC miner would be when the profit of the block generation would not even cover the electricity consumption in the case they won it? otherwise it would be kept on at a static hash rate per ASIC machine with a fixed power consumption. Awesome information. thanks.

This may not always be true. For example miners will sometimes continue to mine if they think that the price of bitcoin will increase in the future.

I believe the transaction fees on the Blockchain will always be higher than on LN because if transactions on the Blockchain becomes cheaper than using LN people would just transact using the Blockchain instead. The inverse would also be true. This would continue until an equilibrium is reached.

[ranochigo]

What do you think about the hash rate supply once the block reward becomes small and transaction fees dominant? I don't see an incentive for end users (tx requester) to place a higher fee if the transaction volume was to drop (e.g. when transactions are executed on the LNN leaving less volume). At this point the miner would need to feasibly supply hashing to the network, but there is no feedback for end-users to increase fees to maintain the security of the network.

or have i missed something?

Given that the block reward would take quite sometime to reach 0, the efficiency of the ASICs would likely increase quite substantially and the electrical prices would decrease. Transactions are still being transacted on chain, albeit only in the opening and closing of channels in the case of lightning network.
If the volume of Bitcoin transactions increases even more, then the mining would still likely be fairly profitable.

I agree, there is some influence that the miner has due to the process of them liquidating coins to run their operation, but the market price determination involves way more factors. What I am more concerned with is the behavior of the network once the coin supply starts to plateau and we have lightened the load of transaction volumes from the blockchain.

Bitcoin's blockchain will still likely remain as the integral part of the system for the rest of its life, as well as the miners. Lightning network helps with the micro transactions but ultimately, it would all still involve Blockchain. The transaction volume now is quite substantial, if Lightning network becomes a reality, then normal day-to-day transactions would be feasible and hence a higher overall transaction volume.