The deflationary problem

[Sweft]

So your saying you expect it to stop?  I find little evidence that it will ever stop, the continual halfing every 4 years of mining rewards clearly points to a about ~20% deflation per year over the long term as the price of BTC must rise to keep mining profitable.

Not necessarily. Fees can rise, as well.

Once we get closer to market saturation, and have a better developed economy - by which I mean more goods and services valued directly in bitcoin - we'll see a lot smoother price movements.

I already addressed this issue, please read the thread.

Fees are capped at 100% so increasing fees cannot support exponential growth.

If bitcoin is still experiencing exponential growth in 2140, the world will have much greater problems than securing the bitcoin network, because that will require many times the current world population and that the population also be growing exponentially.

Network hash must grow at least at the rate of Moore's Law.  This is also known as Sweft's Law.  If Bitcoin fails to satisfy this condition, then it will be prone to attack.

Because if Network hash rate doesn't at least keep up with computing power, it can be outcompeted, and lose the 51% battle?

You seem to have forgotten the second half of Moore's law. Hardware also gets cheaper as it gets more powerful. Fees will more than keep miners in the latest hardware.

I think you are one step closer to realizing the problem.

Hash must always grow.  

The inclination of early Bitcoin theorists was that Hash would fall if miners did not profit, and the mechanism would regulate itself until miners profited.  This is true for mining profit and always will be.

But if miners do not profit, and the difficulty must come down for them to profit, then the whole network can potentially become compromised.

Thus, as i originally stated hash must increase in accordance with Moore's Law, as well as difficulty, and the protocol should reward mining profit through inflation rather than punish it with deflation.

[Adrian-x]

Gold has been inflating for thousands of years.

I am not defending Bitcoin I see flaws too, to your point, gold is a finite resource, and it inflates very slowly over thousands of years, energy and environmental cost of gold inflation dwarfs that of Bitcoin. Bitcoin is finite and inflates (if only it would inflate like gold) the cost of Bitcoin inflation is insignificant in when compared to gold.

[minerva]

https://bitcointalk.org/index.php?topic=174378.0

I've posted an equation that relies on declining rates of return on mining, allowing increased mining to produce more hashes, but also for more coins to be produced.

[myrkul]

Because if Network hash rate doesn't at least keep up with computing power, it can be outcompeted, and lose the 51% battle?

You seem to have forgotten the second half of Moore's law. Hardware also gets cheaper as it gets more powerful. Fees will more than keep miners in the latest hardware.

I think you are one step closer to realizing the problem.

Hash must always grow.  

Nope, it must not. In order for the network to be secure, 51% of hashing power must remain in "white-hat" hands.

Even if a 51% attack happened, it wouldn't do much, nor last long (as the other miners could simply buy cheap hardware and knock him back under 51%).

It's more profitable to play by the rules, than to try and subvert them.
See: https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_computing_power

[Sweft]

Bump.

[JoelKatz]

Even if a 51% attack happened, it wouldn't do much, nor last long (as the other miners could simply buy cheap hardware and knock him back under 51%).

They could, but would they? If they feared the 51% would continue, they'd have no assurance of any return on their investment. If the price falls, some existing miners might be discouraged and stop mining, possibly faster than new mining could be added, especially when it wasn't profitable in the short term.

It's more profitable to play by the rules, than to try and subvert them.
See: https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_computing_power

That assumes that their profit comes from Bitcoin rather than a competitor (which could even be conventional fiat currency). Consider, for example, a future time when Bitcoin prices are high but the block reward is low enough that mining isn't all that profitable. If someone shorts Bitcoins (perhaps even by investing in a competitor, which could even be something like Western Union, PayPal, or Visa), they might actually find it profitable to launch a 51% attack.

One thing that helps Bitcoin resist this kind of attack is that it can't be mined efficiently with general purpose hardware. This means that miners probably won't stop mining at the drop of a hat because they can't sell their mining hardware to someone who is going to do something else with it anyway or use it for something else -- they will have to try to keep Bitcoin healthy event if that means short term losses otherwise their expensive mining hardware is just scrap metal. And it also means that an attacker has to invest in hardware that they can *only* use to mine Bitcoins. They can't rent general purpose hardware to attack the network at a low cost. Coins that use ASIC-unfriendly mining schemes are much more vulnerable to this kind of attack than Bitcoin is.

[myrkul]

Even if a 51% attack happened, it wouldn't do much, nor last long (as the other miners could simply buy cheap hardware and knock him back under 51%).

They could, but would they? If they feared the 51% would continue, they'd have no assurance of any return on their investment. If the price falls, some existing miners might be discouraged and stop mining, possibly faster than new mining could be added, especially when it wasn't profitable in the short term.

It's more profitable to play by the rules, than to try and subvert them.
See: https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_computing_power

That assumes that their profit comes from Bitcoin rather than a competitor (which could even be conventional fiat currency). Consider, for example, a future time when Bitcoin prices are high but the block reward is low enough that mining isn't all that profitable. If someone shorts Bitcoins (perhaps even by investing in a competitor, which could even be something like Western Union, PayPal, or Visa), they might actually find it profitable to launch a 51% attack.

I can't refute those points, but the next one makes them all sort of moot.

One thing that helps Bitcoin resist this kind of attack is that it can't be mined efficiently with general purpose hardware. This means that miners probably won't stop mining at the drop of a hat because they can't sell their mining hardware to someone who is going to do something else with it anyway or use it for something else -- they will have to try to keep Bitcoin healthy event if that means short term losses otherwise their expensive mining hardware is just scrap metal. And it also means that an attacker has to invest in hardware that they can *only* use to mine Bitcoins. They can't rent general purpose hardware to attack the network at a low cost. Coins that use ASIC-unfriendly mining schemes are much more vulnerable to this kind of attack than Bitcoin is.

[Sweft]

Even if a 51% attack happened, it wouldn't do much, nor last long (as the other miners could simply buy cheap hardware and knock him back under 51%).

They could, but would they? If they feared the 51% would continue, they'd have no assurance of any return on their investment. If the price falls, some existing miners might be discouraged and stop mining, possibly faster than new mining could be added, especially when it wasn't profitable in the short term.

It's more profitable to play by the rules, than to try and subvert them.
See: https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_computing_power

That assumes that their profit comes from Bitcoin rather than a competitor (which could even be conventional fiat currency). Consider, for example, a future time when Bitcoin prices are high but the block reward is low enough that mining isn't all that profitable. If someone shorts Bitcoins (perhaps even by investing in a competitor, which could even be something like Western Union, PayPal, or Visa), they might actually find it profitable to launch a 51% attack.

One thing that helps Bitcoin resist this kind of attack is that it can't be mined efficiently with general purpose hardware. This means that miners probably won't stop mining at the drop of a hat because they can't sell their mining hardware to someone who is going to do something else with it anyway or use it for something else -- they will have to try to keep Bitcoin healthy event if that means short term losses otherwise their expensive mining hardware is just scrap metal. And it also means that an attacker has to invest in hardware that they can *only* use to mine Bitcoins. They can't rent general purpose hardware to attack the network at a low cost. Coins that use ASIC-unfriendly mining schemes are much more vulnerable to this kind of attack than Bitcoin is.

ASIC's are tremendously beneficial to the future of cryptocurrencies.  The fact that bitcoin lasted long enough to herald in ASICs is a great accomplishment for all future cryptocurrencies that will compete on the basis of which economic protocol is most efficient.

That said, once the hashrate matures we cannot simply rely on the fact that ASIC miners will not stop mining.  The hashrate must double every two years, otherwise the monetary barrier to a double spend become  smaller and smaller.

[JoelKatz]

I can't refute those points, but the next one makes them all sort of moot.

I hope so. I think it's worth working on ways to reduce the probability and severity of this attack. So, in an academic sense, it's well worth worrying about. But it would be really odd to argue that people should avoid Bitcoin because of the very small risk of this kind of attack.

In addition, though the known solutions (at least known to me) to a 51% attack (from a motivated attacker with significant resources) are pretty awful, it's not necessarily the total death of Bitcoin. So while it would be a severe blow (as the block chain split was, as the recent price crash was), I would expect Bitcoin would survive somehow. Bitcoin's death has been predicted so many times and Bitcoin has weathered crises.

[firefop]

I just wanted to wrap this thread up by pointing out that moores law will break down at some point once we can't make smaller silicon... or at least change to a different time frame. Once bitcoin asics are at the practical limits of silicon's molecular size (is it 4 or 5 nano meters?) we're dead in the water on this issue until quantum computing... at which point we'll have to change the proof of work anyway.

Moore said it himself: the following an extracted from this transcript  http://ftp://download.intel.com/museum/Moores_Law/Video-Transcripts/Excepts_A_Conversation_with_Gordon_Moore.pdf  in 2005.

Interviewer: How long can it continue?

Gordon Moore: I think Moore’s Law will continue as long as Moore does anyhow! Ha ha ha... I’m periodically amazed at how we’re able to make progress. Several times along the way, I thought we reached the end of the line, things taper off, and our creative engineers come up with ways around them. I can think of at least 3-4 things that seemed like formidable barriers that we just blew past without any hesitation and in fact, the board meeting we came up I saw pictures that go a couple of generations beyond anything I had seen previously in regards to the highest transistors we can build... so I think we’ve got quite a bit of life yet. I’ve  ever been able to see beyond the next three generations of technology. Three generations of technology is about 6-8 years and I can see that far now, things haven’t really changed. Eventually they’re going to have to. Materials are made of atoms, and we’re getting suspiciously close to some of the atomic dimensions with these new structures, but I’m sure we’ll find ways to squeeze even further than we think we presently can.

[MoonShadow]

In addition, though the known solutions (at least known to me) to a 51% attack (from a motivated attacker with significant resources) are pretty awful, it's not necessarily the total death of Bitcoin. So while it would be a severe blow (as the block chain split was, as the recent price crash was), I would expect Bitcoin would survive somehow. Bitcoin's death has been predicted so many times and Bitcoin has weathered crises.

From my own perspectives, the block chain split was of zero consequence.  If it were not for the activity on this forum at the time, and my interaction with this forum, I (personally) wouldn't have even noticed.  The same would have been true for most users who either don't spend bitcoins on a daily basis or don't fret over the time to confirm.  Once users are competing for blockspace, a 24 hour long time to confirm will become more commonplace for much more mundane reasons.  And the price crash wasn't even the worst crash that I've seen since I've been here.  On a percentage basis, dropping from about $250 to about $50 over the course of a week or two isn't as bad as dropping from $32 to $2 in 2011 (2012?), although it happened much faster.

[JoelKatz]

From my own perspectives, the block chain split was of zero consequence.  If it were not for the activity on this forum at the time, and my interaction with this forum, I (personally) wouldn't have even noticed.  The same would have been true for most users who either don't spend bitcoins on a daily basis or don't fret over the time to confirm.  Once users are competing for blockspace, a 24 hour long time to confirm will become more commonplace for much more mundane reasons.

I think the fact that someone had a transaction with six confirmation that was then undone was extremely significant.

And the price crash wasn't even the worst crash that I've seen since I've been here.  On a percentage basis, dropping from about $250 to about $50 over the course of a week or two isn't as bad as dropping from $32 to $2 in 2011 (2012?), although it happened much faster.

True, but it happened at a time when Bitcoin's rise was big news and there was a high rate of influx of new users. These are the conditions under which you would never expect such a drop. And it happened due to a failure at a choke point whose significant was not, I think, generally recognized.

But I agree with your point. You can see these events as Earth shattering or meaningless. The important thing is that bad things happened and Bitcoin barely noticed. So if you're worried about bad things happening, the evidence suggests that Bitcoin will survive.

[MoonShadow]

From my own perspectives, the block chain split was of zero consequence.  If it were not for the activity on this forum at the time, and my interaction with this forum, I (personally) wouldn't have even noticed.  The same would have been true for most users who either don't spend bitcoins on a daily basis or don't fret over the time to confirm.  Once users are competing for blockspace, a 24 hour long time to confirm will become more commonplace for much more mundane reasons.

I think the fact that someone had a transaction with six confirmation that was then undone was extremely significant.

Well, you have a point there.

And the price crash wasn't even the worst crash that I've seen since I've been here.  On a percentage basis, dropping from about $250 to about $50 over the course of a week or two isn't as bad as dropping from $32 to $2 in 2011 (2012?), although it happened much faster.

True, but it happened at a time when Bitcoin's rise was big news and there was a high rate of influx of new users. These are the conditions under which you would never expect such a drop. And it happened due to a failure at a choke point whose significant was not, I think, generally recognized.

Those are exactly the conditions that I would expect a big drop, due to the eventual popping of a bubble.  And that is largely what happened.  Honestly, I expected it weeks before it happened, and lost a small fortune short selling with a stop-loss limit  at $72.  It never eeven got back down to where I thought it made fundamental sense, so I would have lost it anyway.

But I agree with your point. You can see these events as Earth shattering or meaningless. The important thing is that bad things happened and Bitcoin barely noticed. So if you're worried about bad things happening, the evidence suggests that Bitcoin will survive.

It's much like the point about violence and terrorrism; terror attacks make big news but not really big body counts, as a nominally "free" US citizen is 9 times more likely to die by the unwarrented actions of a police officer than a terrorist.

Really, the perception of risks are out of prooportion with the reality of those same risks.

[Sweft]

From my own perspectives, the block chain split was of zero consequence.  If it were not for the activity on this forum at the time, and my interaction with this forum, I (personally) wouldn't have even noticed.  The same would have been true for most users who either don't spend bitcoins on a daily basis or don't fret over the time to confirm.  Once users are competing for blockspace, a 24 hour long time to confirm will become more commonplace for much more mundane reasons.

I think the fact that someone had a transaction with six confirmation that was then undone was extremely significant.

Well, you have a point there.

And the price crash wasn't even the worst crash that I've seen since I've been here.  On a percentage basis, dropping from about $250 to about $50 over the course of a week or two isn't as bad as dropping from $32 to $2 in 2011 (2012?), although it happened much faster.

True, but it happened at a time when Bitcoin's rise was big news and there was a high rate of influx of new users. These are the conditions under which you would never expect such a drop. And it happened due to a failure at a choke point whose significant was not, I think, generally recognized.

Those are exactly the conditions that I would expect a big drop, due to the eventual popping of a bubble.  And that is largely what happened.  Honestly, I expected it weeks before it happened, and lost a small fortune short selling with a stop-loss limit  at $72.  It never eeven got back down to where I thought it made fundamental sense, so I would have lost it anyway.

But I agree with your point. You can see these events as Earth shattering or meaningless. The important thing is that bad things happened and Bitcoin barely noticed. So if you're worried about bad things happening, the evidence suggests that Bitcoin will survive.

It's much like the point about violence and terrorrism; terror attacks make big news but not really big body counts, as a nominally "free" US citizen is 9 times more likely to die by the unwarrented actions of a police officer than a terrorist.

Really, the perception of risks are out of prooportion with the reality of those same risks.

The issue isn't any one single double spend attack.

The issue is a long-term, ongoing double spend attack where the attacker holds a significant amount of hash.

[townf]

http://www.marketoracle.co.uk/Article39704.html
The author of this article posted it here in this forum for discussion:
https://bitcointalk.org/index.php?topic=160612.0

51% attack can also exist as a "benevolent" mining monopoly or cartel that doesn't necessarily wreck the blockchain or ruin the value of the cryptocurrency, but simply controls the majority of transaction processing. It then controls the flow of money, who gets charged transaction fees, who doesn't, who even gets their transactions processed. They choose what makes it into the blockchain and what doesn't. They could freeze accounts or move coins to other accounts at will, behind the scenes, benevolently of course, to keep us safe or some shit like that, or without the majority of people knowing (or caring probably) that they can do this, as a matter of their policy.

As soon as the block rewards dry up, this cannot help but become a reality. It is a mathematical, economic inevitibilty, because joe shmoe miner can't afford to prevent it without block rewards. He will go out of business, and that is who we depend on to keep the network honest. We need joe shmoe to be at least 51%, not the walmart-amazon-google network hash cartel to be 51%. Don't let joe shmoe go broke mining!

This cartel will be comprised of corporations or entities who can subsidize their own mining operations, as in pay their electric bill out of their own pocket, buy their hardware, etc without needing block rewards or transaction fees. The joe shmoe public mining pools whose members depend on these compensations to operate would dry up, thus surrendering the 51% of network hash and more to fewer and fewer entities, which will then start acting like a cartel or monopoly and dictate their policies with the clout of essentially holding as hostage the blockchain.

Whoever doesn't think this will happen when the incentive to mine is gone is a fool or is too lazy to think into the future.

The economics of deflation (and inflation) is harmful to participants in an economy anyway in my opinion but that is for another thread. That's more of a general opinion on a complex issue, but the sweft law thing is not an opinion, it is bombproof math and economic reality.

[MoonShadow]

From my own perspectives, the block chain split was of zero consequence.  If it were not for the activity on this forum at the time, and my interaction with this forum, I (personally) wouldn't have even noticed.  The same would have been true for most users who either don't spend bitcoins on a daily basis or don't fret over the time to confirm.  Once users are competing for blockspace, a 24 hour long time to confirm will become more commonplace for much more mundane reasons.

I think the fact that someone had a transaction with six confirmation that was then undone was extremely significant.

Well, you have a point there.

And the price crash wasn't even the worst crash that I've seen since I've been here.  On a percentage basis, dropping from about $250 to about $50 over the course of a week or two isn't as bad as dropping from $32 to $2 in 2011 (2012?), although it happened much faster.

True, but it happened at a time when Bitcoin's rise was big news and there was a high rate of influx of new users. These are the conditions under which you would never expect such a drop. And it happened due to a failure at a choke point whose significant was not, I think, generally recognized.

Those are exactly the conditions that I would expect a big drop, due to the eventual popping of a bubble.  And that is largely what happened.  Honestly, I expected it weeks before it happened, and lost a small fortune short selling with a stop-loss limit  at $72.  It never eeven got back down to where I thought it made fundamental sense, so I would have lost it anyway.

But I agree with your point. You can see these events as Earth shattering or meaningless. The important thing is that bad things happened and Bitcoin barely noticed. So if you're worried about bad things happening, the evidence suggests that Bitcoin will survive.

It's much like the point about violence and terrorrism; terror attacks make big news but not really big body counts, as a nominally "free" US citizen is 9 times more likely to die by the unwarrented actions of a police officer than a terrorist.

Really, the perception of risks are out of prooportion with the reality of those same risks.

The issue isn't any one single double spend attack.

The issue is a long-term, ongoing double spend attack where the attacker holds a significant amount of hash.

I do know what a 51% attack is.  Even then, it would be limited to whomever had recently sold the attackers something of value.  A 51% attacker cannot steal coins that he has never legitimately possessed.

[molecular]

Network hash must grow at least at the rate of Moore's Law.  This is also known as Sweft's Law.  If Bitcoin fails to satisfy this condition, then it will be prone to attack.

Why wouldn't hashrate grow according to Moore's law. If hardware performance increases in hashrate per $, everything else being equal, why wouldn't the bitcoin network hashrate increase accordingly and Sweft's Law be satisfied?

There will always be new miners entering the game with new up-to-date hardware (or existing miners upgrading) and old rigs will drop out, keeping bitcoin hashing equipment up-to-speed with moore's law.

[Sweft]

http://www.marketoracle.co.uk/Article39704.html
The author of this article posted it here in this forum for discussion:
https://bitcointalk.org/index.php?topic=160612.0

51% attack can also exist as a "benevolent" mining monopoly or cartel that doesn't necessarily wreck the blockchain or ruin the value of the cryptocurrency, but simply controls the majority of transaction processing. It then controls the flow of money, who gets charged transaction fees, who doesn't, who even gets their transactions processed. They choose what makes it into the blockchain and what doesn't. They could freeze accounts or move coins to other accounts at will, behind the scenes, benevolently of course, to keep us safe or some shit like that, or without the majority of people knowing (or caring probably) that they can do this, as a matter of their policy.

As soon as the block rewards dry up, this cannot help but become a reality. It is a mathematical, economic inevitibilty, because joe shmoe miner can't afford to prevent it without block rewards. He will go out of business, and that is who we depend on to keep the network honest. We need joe shmoe to be at least 51%, not the walmart-amazon-google network hash cartel to be 51%. Don't let joe shmoe go broke mining!

This cartel will be comprised of corporations or entities who can subsidize their own mining operations, as in pay their electric bill out of their own pocket, buy their hardware, etc without needing block rewards or transaction fees. The joe shmoe public mining pools whose members depend on these compensations to operate would dry up, thus surrendering the 51% of network hash and more to fewer and fewer entities, which will then start acting like a cartel or monopoly and dictate their policies with the clout of essentially holding as hostage the blockchain.

Whoever doesn't think this will happen when the incentive to mine is gone is a fool or is too lazy to think into the future.

The economics of deflation (and inflation) is harmful to participants in an economy anyway in my opinion but that is for another thread. That's more of a general opinion on a complex issue, but the sweft law thing is not an opinion, it is bombproof math and economic reality.

This is a good point.  Block reward is fixed, tx fees are not.  Thus, someone can solve blocks at negative cost and bankrupt all other miners, since he can reduce his tx fee to 0.

Inflation also solves this problem because a mining cartel cannot remove all of the income from the solving blocks.

[Sweft]

Network hash must grow at least at the rate of Moore's Law.  This is also known as Sweft's Law.  If Bitcoin fails to satisfy this condition, then it will be prone to attack.

Why wouldn't hashrate grow according to Moore's law. If hardware performance increases in hashrate per $, everything else being equal, why wouldn't the bitcoin network hashrate increase accordingly and Sweft's Law be satisfied?

There will always be new miners entering the game with new up-to-date hardware (or existing miners upgrading) and old rigs will drop out, keeping bitcoin hashing equipment up-to-speed with moore's law.

An explanation of Sweft's Law for those having trouble understanding it.

Moore's Law states that the number of transistors doubles every two years.

This means two years from now, the cost of the network (hash / ($ / hash)), given equivalent hash, will be half.

And half in another two years. And half in another 4 years.

That means in 6 years it will take 1/8th the amount of capital to replicate the original hash.

Thus, the network must grow at the rate of Moore's Law to sustain the value of its hash so that the value to replicate the original hash does not decrease.

That means that the network must double its hash every two years.

Every two years there has to be a capital investment equivalent to the cost of network hash / 2 to double the hash in accordance with Sweft's Law.


But then we also have to calculate electricity costs, which i will do later.

[MoonShadow]

This is a good point.  Block reward is fixed, tx fees are not.  Thus, someone can solve blocks at negative cost and bankrupt all other miners, since he can reduce his tx fee to 0.

This is a nonsense statement.  It's impossible for any entity to bankrupt all other miners, no matter his transaction fees.  Not only would any such miner have to aquire the capital investment in order to build his network to beyond 51%, an uphill battle considering the vast majority of the current running network (now or at any other time) is a sunk cost already paid for; he would also have to compete with the hobbyist miners who benefit from co-generation of waste heat.

I have a GPU based mining rig that I choose not to use, that I can use for space heating in my garage at will.  No miner can bankrupt me.

Therefore your "law" is invalid.