[If tx limit is removed] Disturbingly low future difficulty equilibrium

[Meni Rosenfeld]

If the price of a bitcoin increases, so does the incentive to attack the network

So does the incentive to secure it.

Exactly, that's my point. You can't argue "the current level of security requires spending $10M / year, and that's a small price to pay for a global financial system". By the time Bitcoin is a ubiquitous global financial system, securing it will cost $1T / year, and nobody will cough that up without personal incentives. Which is precisely the "tragedy of the commons" problem we've been talking about.

[1714051776]

1714051776

[1714051776]

1714051776

[1714051776]

1714051776

[cunicula]

It is worth pointing out that the disturbingly low difficulty is just the competitive equilibrium. The monopolistic equilibrium seems much more likely to occur. One agent will take control of 51% of hashing power, prohibit other miners from operating, raise txn fees to a renumerative level, and make sure that the currency operates as normal (e.g. no double-spends and no denial of service). He will earn a lot of money and provide a more secure payments service, benefiting users. This equilibrium seems much more plausible than frequent, isolated double-spend attacks coming from different people. Tragedy of the commons should quickly degenerate into a hostile takeover under a benevolent monopolist who will have a strong motivation to improve the security situation.


Note: Despite this, I still recognize a danger that the monopolist will have some competing outside interest (e.g. he also owns paypal). In this case, he might behave maliciously (double-spending alone is not a plausible reason, shorting isn't likely to be one either [the market won't be deep enough]). Proof-of-stake is the best mechanism for guarding against this possibility.

[ripper234]

It is worth pointing out that the disturbingly low difficulty is just the competitive equilibrium. The monopolistic equilibrium seems much more likely to occur. One agent will take control of 51% of hashing power, prohibit other miners from operating, raise txn fees to a renumerative level, and make sure that the currency operates as normal (e.g. no double-spends and no denial of service). He will earn a lot of money and provide a more secure payments service, benefiting users. This equilibrium seems much more plausible than frequent, isolated double-spend attacks coming from different people. Tragedy of the commons should quickly degenerate into a hostile takeover under a benevolent monopolist who will have a strong motivation to improve the security situation.


Note: Despite this, I still recognize a danger that the monopolist will have some competing outside interest (e.g. he also owns paypal). In this case, he might behave maliciously (double-spending alone is not a plausible reason, shorting isn't likely to be one either [the market won't be deep enough]). Proof-of-stake is the best mechanism for guarding against this possibility.

cunicula, I haven't seen people share in your strong belief in this scenario.

I think that there are so many ideological and practical reasons to oppose a monopoly (even a "benevolent monopoly"), that it's practically obvious to a lot of people that such a monopoly would mean the death of Bitcoin. If this were indeed the case, then forming a (public) monopoly is not profitable endevour, because if one is successful in creating the monopoly, then the value of Bitcoin will plummet.

By these arguments, miners do not have an incentive to form a monopoly, and so the scenario we'll actually end up with is indeed competitive mining. I assume that some solution to the Tragedy of the Commons issue is indeed adopted, be it Proof of Stake or Dominant Assurance Contracts.

[cunicula]

Yes, I agree that forum members are vehemently opposed to a mining monopoly.  I think this is an irrational view which is related to the political leanings of the  current user base. I expect this view to change over time as the user base diversifies. I don't understand how bitcoin could become successful AND not attract a monopolist. There is nothing in the technology to exclude a monopolist from taking over, proof-of-stake helps but it doesn't change this. Monopoly is potentially quite profitable. People may not like this, but their feelings don't affect the situation. Bitcoin use must eventually be supported by utility as a payments system. I don't think users will abandon a useful payments system because a monopolist is verifying their txns rather than a network of competitive miners. Once demand for money transfer becomes the primary reason for use of the technology, people won't care who is verifying the txns as long as they are cheap, quick, and cash-like.  

[Meni Rosenfeld]

Yes, I agree that forum members are vehemently opposed to a mining monopoly.  I think this is an irrational view which is related to the political leanings of the  current user base. I expect this view to change over time as the user base diversifies. I don't understand how bitcoin could become successful AND not attract a monopolist. There is nothing in the technology to exclude a monopolist from taking over, proof-of-stake helps but it doesn't change this. Monopoly is potentially quite profitable. People may not like this, but their feelings don't affect the situation. Bitcoin use must eventually be supported by utility as a payments system. I don't think users will abandon a useful payments system because a monopolist is verifying their txns rather than a network of competitive miners. Once demand for money transfer becomes the primary reason for use of the technology, people won't care who is verifying the txns as long as they are cheap, quick, and cash-like.  

Can you explain how a monopolized Bitcoin is better than a central currency issuer? The entire point of Bitcoin is decentralization, if we don't have that, what do we have left? And why do we need the whole proof of work/stake thing, instead of the monopolist just synchronizing transactions in its internal database?

You could take pretty much any argument in favor of Bitcoin and use it against a monopoly. What if it charges exorbitant fees of $0.3 + 3%? What if its servers are down and nobody can process transactions? What if it is attacked by some government agency? What if it goes bankrupt due to bad management? What if it starts posing restrictions on how one can use their bitcoins?

[da2ce7]

While not directly related to how high or low (I think the difficulty will be low).
I have posted a separate thread on why having a low "base load" hashing power isn't a bad thing... Rarther a good thing:

https://bitcointalk.org/index.php?topic=68218.0

Hope my above thread shows why "difficulty" isn't the be-end-all on bitcoin security.

[cunicula]

I agree that competitive mining has some advantages. However, in other respects monopoly looks superior.  Just as importantly, mining looks like a natural monopoly. If monopolized mining is highly profitable, then bitcoin will likely become a monopoly. The system should be prepared to accommodate a monopolist. Proof-of-stake is a good way of ensuring that the monopolist will behave himself.

Can you explain how a monopolized Bitcoin is better than a central currency issuer? And why do we need the whole proof of work/stake thing, instead of the monopolist just synchronizing transactions in its internal database?

Transparent, psuedonymous, and irreversible txns. It is not clear how this can be maintained without an alternate currency system such as bitcoin. A company which owns the majority of bitcoins and maintains a centralized txn database is more trustworthy than a company which just maintains a centralized txn database.

What if it charges exorbitant fees of $0.3 + 3%?

The monopolist will have lower marginal costs because most of its hashing capacity could remain idle. Fees could go up or down. I think they would go down because the marginal cost savings are potentially extremely large, but it is an empirical question. Remember that implicit fees are currently outrageously high. Low fees should not to be a talking point when so much new money is being issued.

What if its servers are down and nobody can process transactions?

I think they would have strong incentives to keep their servers up. If they went down for a prolonged period, then private miners could take over. The monopolist would have strong incentives to recognize their txns.

What if it is attacked by some government agency?

Proof-of-work is just as easy to attack under centralization as it is under decentralization. If the monopolist is unable to maintain dominant hashing power, then  competitive miners or another monopolist could take over. If the gov't decides to maintain dominant hashing power, then competitive miners would be fucked anyway. Same goes for proof-of-stake, if a private actor can silently acquire a majority stake, then a gov't could do so easily.

What if it goes bankrupt due to bad management?

If it ceases hashing, then competitive miners or another monopolist will take over. The blockchain will remain transparent and there will be a market for control over it.

What if it starts posing restrictions on how one can use their bitcoins?

If it does this voluntarily, it will lose a principal niche market for drugs, money laundering, etc. I don't think it would want to do this. If the government forced the monopolist to do this, then either a) another monopolist will take over who can operate more profitably in a jurisdiction without these constraints, or b) the gov't would act the same way against a p2p chain. In any case, it is not plausible for bitcoin to grow tremendously and simultaneously remain unregulated.

[rjk]

All this talk of a monopoly. Do you happen to work for Microsft, Adobe, the US Government, etc?

[MoonShadow]

I agree that competitive mining has some advantages. However, in other respects monopoly looks superior.  Just as importantly, mining looks like a natural monopoly.

Natural monopolies are not monopolies, but temporarily dominant economic players. 

[FreeMoney]

All this talk of a monopoly. Do you happen to work for Microsft, Adobe, the US Government, Hasbro etc?

missed one :p

[rjk]

All this talk of a monopoly. Do you happen to work for Microsft, Adobe, the US Government, Hasbro etc?

missed one :p

QFT

[kjj]

The monopolist will have lower marginal costs because most of its hashing capacity could remain idle. Fees could go up or down. I think they would go down because the marginal cost savings are potentially extremely large, but it is an empirical question. Remember that implicit fees are currently outrageously high. Low fees should not to be a talking point when so much new money is being issued.

Wait a second.  You think that the monopoly miner will be able to leave most of their hashing power offline, as a fleet-in-being?  Won't that just make the scaling effects worse?

[cunicula]

Wait a second.  You think that the monopoly miner will be able to leave most of their hashing power offline, as a fleet-in-being?  Won't that just make the scaling effects worse?

Are you describing the increasing-returns to scale as the scaling problem? If so, then yes it makes the scaling problem worse.

If the monopolist has say 55%, then he can reverse blocks made by other miners. Since other miners know that the monopolist plans to fire up his 55% if they try to mine, private miners have no incentive to keep mining. Therefore, private miners will take all there rigs offline. The monopolist doesn't need to keep his rigs running, he just needs to be able to credibly threaten to start them and recapture all privately mined blocks. Therefore, the monopolist can get away with powering just  2% of the networks hashing capacity and keeping the other 53% in reserve in case he faces an attempt at a competition. Like private miners, he will have a lot of capital costs (though less than they do), but he will have minimal electricity cost. This is likely a source of tremendous cost savings (and lower txn fees) under monopoly.

As a monopolist, it will make sense to pick the highest MH/$ solution for the vast majority of rigs. MH/J won't be important to him.

Proof-of-stake doesn't suffer from this additional scaling issue because everything is loaded into capital costs and electricity costs are unimportant.

[cunicula]

It seems to me that a monopolist would have the most incentive to create a monopoly today. Why isn't there a monopoly on mining then? What changes in the future that gives a monopolist more incentive to create a monopoly? A lower block reward along with an assumed lower barrier to gain a monopoly?

I believe it is just the risk. Most people think bitcoin is a joke. No one is willing to plunk down 10 million dollars on hardware to hash joke currency. I personally would need 100 million in net worth or more to consider pursuing this individually. You could potentially lose much of the investment if monopolizing the currency causes the bubble to pop. There is a lot of legal risk too. If bitcoin is a proven technology and its valuation is supported by actual use as opposed to an expected possibility of future use, then it will look less like a joke, less risky, and more legal. Moreover, valuation will be based on fundamentals rather than emotional issues such as whether there is just one miner or a multitude of miners. In this case, people or a corporation will be willing to plunk down the money.

[markm]

It seems to me that a monopolist would have the most incentive to create a monopoly today. Why isn't there a monopoly on mining then? What changes in the future that gives a monopolist more incentive to create a monopoly? A lower block reward along with an assumed lower barrier to gain a monopoly?

I believe it is just the risk. Most people think bitcoin is a joke. No one is willing to plunk down 10 million dollars on hardware to hash joke currency. You could potentially lose much of the investment if monopolizing the currency causes the bubble to pop. There is a lot of legal risk too. If bitcoin is a proven technology and its valuation is supported by actual use as opposed to an expected possibility of future use, then it will look less like a joke and less risky. Moreover, valuation will be less sensitive to emotional issues such as whether there is just one miner or a multitude of miners. In this case, people will be willing to plunk down the money.

Plunk it down on what, exactly? CPUs? GPUs? FPGAs? ASICs?

Or will they cover the bases, so that moving our own wealth out of bitcoins into oh lets pick litecoins out of our asses to start with and see if someone is a criminal deliberately attempting to disrupt other people's networking and data recording - a computing or data crime of some sort maybe in some jurisdictions - or an honest investor honestly interested in helping us achieve the objectives of our many and various branches of the cryptocurrency movement by supporting and enhancing the value and security of everyon'es cryptocurrency holdings won't help us elude their grasping grasp?

-MarkM-

[cunicula]

All this talk of a monopoly. Do you happen to work for Microsft, Adobe, the US Government, etc?

Citibank. Just kidding, but some of my classmates went to work for them.

[kjj]

Wait a second.  You think that the monopoly miner will be able to leave most of their hashing power offline, as a fleet-in-being?  Won't that just make the scaling effects worse?

Are you describing the increasing-returns to scale as the scaling problem? If so, then yes it makes the scaling problem worse.

No, I'm talking about the increasing-overhead to scale problem.  Or, at least the apparent overhead.

The bigger your mining operation is, the harder it is to ignore your overhead.  A couple hundred megahashes per second requires no infrastructure, since it is just a video card or two that you already own, and a small fraction of your monthly power and network bills.  A couple of terahashes per second is going to need its own building(s) with dedicated generators and multiple internet connections.

[jtimon]

Proportional to the total ever minted doesn't sound right either; proportional to number actively circulating sounds more apropos maybe. Maybe plus dark pools, but maybe activity can help guess which are lost and which are merely dark pools that have not made any moves in the last few generations or years or centuries or whatever.

-MarkM-

EDIT: Remember, I am pointing at aux chains, increased reward can arise by increased number of high difficulty chains. Whereas you seem to be imagining your proposed chain is the be all and end all, hogging all the hashing power to itself, refusing to share the bountiful potential of merged mining...

No I'm just curious about how some of the calculations people hav made here would apply to freicoin. Mainly because there's some people who think 5% annual demurrage is too much for giving it all to miners.
I think is probably too little, but I don't really have a method to calculate the security requirements.

I wanted to post there something like "According to XXXX from the bitcoin forum, 3 (or 4, or 5, I don't know) freicoin chains at 5% demurrrage (or one at 15%, 20%, etc) would be necessary for maintaining the needed security (without initial subsidy, tx fees nor merged mining with other chains)".

Seriously, guys, what's your method for estimating security requirements?

Bitcoiners sound much like all the prervious entrenched monopolists sometimes. Oh gosh we cannot allow hashing power to be spread out, we have to let it go to waste to maintain our monopoly...

I thought breaking monetary monopolies was supposed to be a good thing, that bitcoin is here to help accomplish...

Completely agree.
Bitcoin doesn't need to be protected from altchains. Let the free market work!
Let merged mining save us all (including bitcoin) resources!

[cunicula]

No I'm just curious about how some of the calculations people hav made here would apply to freicoin. Mainly because there's some people who think 5% annual demurrage is too much for giving it all to miners.
I think is probably too little, but I don't really have a method to calculate the security requirements.

To sustain a similar level of security to that currently enjoyed by bitcoin, you would need 30.6% demurrage per annum. Similarly, bitcoin would need to generate txn tax revenue equivalent to 30.6% of demurrage per annum. Fee rates like these are incompatible with successful competition with existing technologies such as paypal. Good luck. You might consider some proof-of-stake system incorporating modest demurrage, since then you could avoid high fees.  I have no strong objection to demurrage.  In fact, since holding vast amounts of wealth in the coin creates an externality (it imposes increased security risks on other users), there is a strong theoretical argument for some modest level of demurrage in a proof-of-work or partial proof-of-work system.

I just dislike the idea of currencies that are not likely to be competitive with existing technologies except for niche uses.

[Boussac]

Future difficulty equilibrium will be found based on three variables (among others):
1/number of miners (ie units of hashing power) mining for bitcoin transactions fees
2/number of miners mining for value-add transaction fees that one can charge to a merchant by providing value-added services (like escrow, third-party signatures, exchange rate hedging, etc.)
3/ number of miners mining for other reasons (e.g politically motivated or holding big stash of coins)

Today it's fair to say that virtually all miners fall in categories 1 or 3.
I believe in a not so distant future, category 2 will grow as well and should be factored in whatever model is put together to project the difficulty equilibrium (I assume there is one non-zero equilibrium)


I havent seen the model yet and have no time now to work on that but I would appreciate inputs from modelling experts.