Author

Topic: pegging cryptocoins to energy costs (Read 1622 times)

legendary
Activity: 1022
Merit: 1033
September 08, 2012, 05:45:59 AM
#11
This is not accurate. Prices are determined by supply and demand. Mining provides the supply, and the economy provides the demand. You don't know what happens to the price unless you know what happens to both supply and demand. In this case, what you wrote is only true if the demand for the coin increases at a constant rate equal to the average mining rate.

Yes, this is not accurate because it is just an illustration, not a concrete description of a process.

If you want to consider it in a semi-formalized fashion, consider prices which would form on futures exchange for these electro-coins. (I assume that prices of electric energy is fixed and number of hashes per unit of energy is also fixed.)

Suppose that delivery date is one month from the point where trading stops. Let's introduce two parameters: B is maximum possible demand for coins at prices above electricity costs (you can also consider a stochastic model: e.g. probability that demand is higher than B is very small), and P is peak hashing rate: maximum amount of coins miners can produce within a month (it is limited by electricity and hashing power available to miners, as well as 'ramp up time').

If B < P, there is no reason for coins to trade significantly above energy costs. (Assuming arbitrage-free market, perfect competition etc.)

If B > P then we can get coins traded above energy costs, but is this a realistic scenario? Moreover, if B(x) is maximal cumulative demand over x time units (e.g. months), I expect that B(x)/x goes down for higher x, i.e. we can come up with tighter bounds, on the other hand P(x)/x goes up with higher x because miners can ramp up hashing power if there is demand. So eventually for some x they will meet.

If we know that price stabilizes in the long term, it creates arbitrage opportunities and so price will stabilize even in short term.

Now let's consider a situation where current price is below electricity costs. Then there is no mining. If we assume that there is coin loss proportional to economic activity (I already mentioned it), at some point supply will fall below demand.

(OK I guess Etlase2 now has much better solution, but I believe even simple rules can significantly stabilize price.)
hero member
Activity: 798
Merit: 1000
September 08, 2012, 03:37:37 AM
#10
I would also like to point out that there is an equilibrium between the cost of mining and the value of the coin. If the cost of mining a coin is greater than the value of the coin, then there will be fewer miners, lowering the difficulty to the point where it is again cost-effective to mine.

As I mentioned in the post above yours, the proposals for Encoin and Decrits both do not allow for the difficulty to decrease. A big part of making this possible was to separate the security of the network from mining. *Mining is not required to secure a cryptocurrency*.

In order to maintain a constant value of your coin you must be able to both add and remove coins from the economy. Now, you might be able to come up with an automatic system that can do that in theory, but it will be close to impossible in practice.

Removing coins from the economy was initially part of the Encoin proposals by destroying unrefunded transaction fees (those not returned via proof-of-economic-activity). Fairly automatic, but also unnecessary when security and mining are separate. Rather than focusing on extreme price stability, I'd rather opt to let people keep their money while everyone knows that the coins generally cost more to produce than they are currently worth, so as long as confidence has not been lost, the currency will eventually return to a point where the value of a coin oscillates around its cost to produce.

Of course, coin backed by energy would have much higher exchange rate fluctuations than USD, something like price fluctuations of base commodities, i.e. 100% change within a year is quite possible.

But still we don't have same situation as with Bitcoin where it can shoot up 1000x.

On that note, an energy-related (rather than "backed") currency would also have a tangibly inherent and quantifiable value basis. Bitcoin only has value based on its demand and the unquantifiable amount of people who feel like selling and in what amounts. Everyone could easily see that it would take X amount of hours with $X amount of hardware and electricity to create X amount of Decrits. There is no such equation possible for Bitcoin. This means the currency could thrive even without central exchanges.
legendary
Activity: 1022
Merit: 1033
September 08, 2012, 12:33:58 AM
#9
In order to maintain a constant value of your coin you must be able to both add and remove coins from the economy. Now, you might be able to come up with an automatic system that can do that in theory, but it will be close to impossible in practice.

That's easy: destructive transaction fees. I.e. transaction fee must be included, but it doesn't go to miner. Thus we assume that there is a loss proportional to coin's "GDP".

Quote
Consider that this is exactly what the Federal Reserve is supposed to do -- make the value the dollar constant. Not only have they never come up with an automatic system that works, but they have consistently done a poor job trying to maintain it manually.

For some values of 'poor'. Fluctuations of USD exchange rate are usually small, something like 5% per year.

On the other hand, Bitcoin exchange rate went from 0.01 to 10 USD within one year (IIRC), so we have 100000% change.

That's quite a difference, right?

Three orders of magnitude vs zero.

Of course, coin backed by energy would have much higher exchange rate fluctuations than USD, something like price fluctuations of base commodities, i.e. 100% change within a year is quite possible.

But still we don't have same situation as with Bitcoin where it can shoot up 1000x.

Quote
For more information on the mathematics of it all, look up dynamical systems and numerical integration.

Thank you, I just have M.Sc. in applied math, what do I know about mathematics?

My description was hand-wavy because I consider all details kinda obvious, not because I don't know math, lol

I suspect you're focusing too much on mining and ignore economic aspects of this: when participants KNOW that coin is pegged, and they KNOW that pegging mechanism works in long term under broad assumptions, price will be stabilized through arbitrage, not through mining.

I think I mentioned this, do you know what arbitrage means?
legendary
Activity: 4466
Merit: 3391
September 08, 2012, 12:22:41 AM
#8
In order to maintain a constant value of your coin you must be able to both add and remove coins from the economy. Now, you might be able to come up with an automatic system that can do that in theory, but it will be close to impossible in practice.

Consider that this is exactly what the Federal Reserve is supposed to do -- make the value the dollar constant. Not only have they never come up with an automatic system that works, but they have consistently done a poor job trying to maintain it manually.

For more information on the mathematics of it all, look up dynamical systems and numerical integration.
legendary
Activity: 4466
Merit: 3391
September 07, 2012, 11:42:13 PM
#7
Suppose one block gives you 1 coin.

  • 1. If electricity it costs to mine one block costs more than 1 coin people will stop mining coins as they will be losing money. Less coins are mined each day=> coin value goes up.
  • 2. If electricity it costs to mine one block costs less than 1 coin people will start mining as many coins as possible because it is profitable. More coins are mined each day => coin value goes down.


This is not accurate. Prices are determined by supply and demand. Mining provides the supply, and the economy provides the demand. You don't know what happens to the price unless you know what happens to both supply and demand. In this case, what you wrote is only true if the demand for the coin increases at a constant rate equal to the average mining rate.

I would also like to point out that there is an equilibrium between the cost of mining and the value of the coin. If the cost of mining a coin is greater than the value of the coin, then there will be fewer miners, lowering the difficulty to the point where it is again cost-effective to mine. If the costs are much lower than the value of the coin, then there will be more miners, which will raise the difficulty and the cost. This means that the cost of mining a coin will always tend be about the same as the value of the coin. Again this doesn't directly affect the value of the coin because it only deals with the supply side.
hero member
Activity: 798
Merit: 1000
September 07, 2012, 06:35:18 PM
#6
Solidcoin's peg is a joke and is just whatever RealSolid decided was an approximation of Koomey's Law (amount of work per joule doubles every 1.5 years), not one based on any real evidence from mining. Unfortunately this is an epic fail as can be seen by the jump from CPU->GPU, GPU->FPGA, FPGA->ASIC. It just doesn't work this way in reality, especially when it comes to mining.

Killerstorm, if you're reading a document then it is most certainly an outdated version of the Encoin proposals. The most up to date version is here: http://justinbporter.com/encoin/doku.php

Some of the problems with using an energy-based currency is that the difficulty can be manipulated (difficulty doesn't go down was a big lightbulb moment), so there have to be some controls, as described here: http://justinbporter.com/encoin/doku.php?id=mint_blocks and here: http://justinbporter.com/encoin/doku.php?id=attacks

Moore's Law adjustment here: http://justinbporter.com/encoin/doku.php?id=moore_s_law Koomey's law adjustment here: http://justinbporter.com/encoin/doku.php?id=koomey_s_law

Decrits uses a much more fine-tuned idea of adjusting the coin award after the difficulty increase, and making it difficult to flood the network with high-output by locking each minter into groups with others and having to wait for some of them to finish and so on. Again they are controls against intentionally raising the difficulty, a big factor when the difficulty doesn't go down.

There is no block chain in the bitcoin sense in either Encoin or Decrits, although Encoin may have sort of had one in the earlier revisions before I more finely tuned things. This was a proposal for a bitcoin-like stable value currency which had to have a form of demurrage to keep the value stable: https://bitcointalksearch.org/topic/new-musings-for-a-stable-currency-64637

But I think that Decrits is the best incarnation of the ideas, as it relies on the least amount of energy waste by giving lots of coins away for free, thus allowing any change in efficiency to be quickly absorbed by the currency and providing the least amount of upset possible. It would also require some significant amount of the network's GDP to be mined before being able to maliciously increase the difficulty while profiting as little as absolutely possible. Encoin's transaction fees were ridiculous to encourage Proof-of-Economic-Activity as security for the network (whereas Decrits goes by the much simpler and cheaper proof-of-share/stake); it would also be more data heavy and less anonymous than Decrits because of that.

Decrits would still be strongly correlated with the cost of electricity plus the cost of hardware (or less the sunk costs of GPUs), which is why I dropped the energy correlation because hardware cost is a significant factor, especially as efficiency increases, so I didn't think there was any point in keeping up with the idea of it being based solely around energy costs.

But I strongly believe Decrits is the true answer to a cryptocurrency rather than backpedaling to a gold-like standard and all of the issues that went along with that. Why would anyone accept fractional reserve when by not doing so, demand for currency will rise and new money will be freely given away to the people rather than the banks? Encoin mining would still be a significant choke hold any time the market expanded by requiring 10x the mining et al to achieve the same effect as Decrits. New merchants in Decrits would be rewarded with free coins simply by transacting; certainly a nice incentive. Everyone who uses the network gets to profit on an expansion without forcing a large chunk of that profit into the hands of GPU manufacturers or electric companies, and not just those that hold lots of currency or those that mine.
legendary
Activity: 1205
Merit: 1010
September 07, 2012, 02:51:21 PM
#5

Yes. We can have either block spacing or we can have a mostly stable exchange rate.

I don't know why block spacing needs to be stable, maybe you do?

IIRC we even have a proposal to have a dynamic "block spacing" on purpose.


Well sure at least you don't want to end up with millions of blocks per second in case your coin gets really popular do you? Nor is it preferable to have no blocks for days if miners gave up?

SolidCoin has been trying to 'peg' to electricity cost since last version afaik, albeit not with constant work per block which obviously is not gonna work in my opinion.

legendary
Activity: 1022
Merit: 1033
September 07, 2012, 02:37:11 PM
#4
is proved by Market.

Also the fact that ppcoin is alive means that market doesn't know a shit Smiley
legendary
Activity: 1022
Merit: 1033
September 07, 2012, 02:36:35 PM
#3
If you make work constant you cannot expect block spacing (10 minutes target) be stable.

Yes. We can have either block spacing or we can have a mostly stable exchange rate.

I don't know why block spacing needs to be stable, maybe you do?

IIRC we even have a proposal to have a dynamic "block spacing" on purpose.

Quote
But of course solidcoin has been trying this for a while.

Do you mean that SolidCoin had a constant work per block?

Quote
You make up a formula regarding Moore's Law and still inflation could end up pretty wild in the event of a mining technology upgrade such as cpu->gpu->asic.

Memory-hard hash functions largely remove GPUs and ASICs from consideration. Although I agree it won't be completely stable, just more stable than currency not pegged to anything.

Quote
One and half years ago I was rejecting Bitcoin based on the belief that an energy backed version would be much natural (more like gold) but I changed my position. Satoshi's variable difficulty and fixed inflation is a better design and is proved by Market.

There was no competition of designs, as far as I know. At least, not pure designs. It also depends on implementations and communities.

If somebody invents an ObviouslyBetterCoin tomorrow, people won't simply jump to it: they already have infrastructure around Bitcoin.
legendary
Activity: 1205
Merit: 1010
September 07, 2012, 02:27:29 PM
#2
If you make work constant you cannot expect block spacing (10 minutes target) be stable. You need to make block pay to be proportional to work. But of course solidcoin has been trying this for a while. You make up a formula regarding Moore's Law and still inflation could end up pretty wild in the event of a mining technology upgrade such as cpu->gpu->asic.

One and half years ago I was rejecting Bitcoin based on the belief that an energy backed version would be much natural (more like gold) but I changed my position. Satoshi's variable difficulty and fixed inflation is a better design and is proved by Market.
legendary
Activity: 1022
Merit: 1033
September 07, 2012, 02:08:28 PM
#1
Recently I've ready Etlase2's Encoin proposal. Description of mining in this system spans several pages and I'm not sure I've got it correctly, but doesn't it boil it down to this:

To make a coin with exchange rate pegged to electricity costs we need to make amount of work per block and bounty constant.

Why it works?

Suppose one block gives you 1 coin.

  • 1. If electricity it costs to mine one block costs more than 1 coin people will stop mining coins as they will be losing money. Less coins are mined each day=> coin value goes up.
  • 2. If electricity it costs to mine one block costs less than 1 coin people will start mining as many coins as possible because it is profitable. More coins are mined each day => coin value goes down.

So we see that exchange rates goes into a direction which aims to compensate difference between coin value and electricity costs.

Certainly, there might be a large lag and large fluctuations since demand isn't constant. BUT we know that in the long run coin value will go electricity costs, so people will see arbitrage opportunities if there is an imbalance, and arbitrage again moves price into the right direction. Currency exchange arbitrage likely works faster than mining "arbitrage".

What's about protection against double-spends? Well, if our only goal is more-or-less stable exchange rate, we can just use merged-mining with Bitcoin for security and use mining only for distribution of coins.

Encoin describes rather elaborate scheme to which allows node to reduce amount of work per block (also reducing bounty), but as I understand it is necessary only for keeping some hash rate when coins are too cheap.

Am I missing something or is it really that simple? Then why don't we have an alt-chain pegged to electricity costs?
Jump to: