Bitcoin & Tragedy of the Commons

[phelix]

Good points by cunicula. Overall I've come to the conclusion that this issue could prove quite challenging. The good thing is that there are many years to work on this, currently our mining network (and thus security) is oversized compared to the size of the market. It will take a long time until this starts to be a real problem but it's good that people are thinking about it already.

yes, but the earlier we find a good solution the better. it should be first tested in a fork, too. Also I am interested if it can improve Namecoin, where the block reward is getting quite low in fiat terms.


I repeat my suggestion from above: Hard code a maximum spread factor of four between the highest and lowest transaction fee within a single block.

Miners will usually want to include the highest fees.
Low fee TXs will sum up and at some point become profitable to include in a low fee block (that does not contain any high fee TXs).

Advantage: The "fee market" will become less steep (compare DeathAndTaxes' post above)

Possible disadvantage: fees of certain height might be delayed because they are too high.

[cbeast]

If the king owns it all the it isn't a tragedy of the commons.

that's why I added the "/sarcasm" 
Wow, if you didn't see that I guess I should not presume everyone has read/seen any Robin Hood stories.

[Rassah]

Sorry for off-topic:
The irony of Robin Hood is that it's a story that demonizes the state, which is literally stealing people's livelyhood through taxes, and shows as heroes the group that takes down said state and returns wealth to its actual owners, the people who worked to produce it... while we commonly think of it as a story that demonizes the rich, who somehow have all this money that shouldn't belong to them, and shows as heroes the group that "steals from the rich and gives to the poor," because the poor deserve that money by virtue of being poor and needing it more.
How the hell did that happen???

D&T - completely agree regarding that type of monopoly. Monopoly through trade secrets is a definite possibility/inevitability. I don't know why cunnicula was bringing up monopoly through any returns to scale, since, yeah, there are minor advantages as you mentioned, but they're minor.

[DeathAndTaxes]

I think (although I can't speak for him) the fact is that any monopoly through superior technology requires a lot of scale.

YOU could today make an ASIC miner which is not only cheaper per MH/s than anything on the market it is also more power efficient than anything on the market (back on napkin is ~500MH/s per watt).  This could be done TODAY... only the minor roadblock of needing ~$10M in capital and the risk that capital would be taking.

[cbeast]

Sorry for off-topic:
The irony of Robin Hood is that it's a story that demonizes the state, which is literally stealing people's livelyhood through taxes, and shows as heroes the group that takes down said state and returns wealth to its actual owners, the people who worked to produce it... while we commonly think of it as a story that demonizes the rich, who somehow have all this money that shouldn't belong to them, and shows as heroes the group that "steals from the rich and gives to the poor," because the poor deserve that money by virtue of being poor and needing it more.
How the hell did that happen???

You are not off topic at all. Tragedy of the Commons is as much a fallacy as The Lorax, Robin Hood, or even applying thermodynamic laws inappropriately when it comes to complex and open systems. Trying to solve a problem that doesn't exist doesn't mean we're redefining the problem (i.e. demonizing the state because the Sheriff < Richard), I am simply saying that applying a Tragedy of the Commons argument to an artificial structure like money is silly. Tragedy of the Commons can only apply to natural systems like Lotka–Volterra equations. Humans are resourceful enough to find technological solutions to resource issues, not that they always do, but that's another issue entirely NOT related to the resources themselves.

[cunicula]

I think (although I can't speak for him) the fact is that any monopoly through superior technology requires a lot of scale.

Yes, D&T is correct as usual.

YOU could today make an ASIC miner which is not only cheaper per MH/s than anything on the market it is also more power efficient than anything on the market (back on napkin is ~500MH/s per watt).  This could be done TODAY... only the minor roadblock of needing ~$10M in capital and the risk that capital would be taking.

You might be able to recoup a lot of that. Just sell as many patented units on the open market until it is just about saturated at a price equal to say four times your marginal cost. I think you could pawn off several million dollars worth. Now make another, larger batch of units and use these to seize control of 51% of the network, expropriate your customer's mining rights, and earn ~35k USD of revenue a day.

I agree that it is still highly risky, but the look on your customers faces would be priceless.

[Phinnaeus Gage]

Sorry for off-topic:
The irony of Robin Hood is that it's a story that demonizes the state, which is literally stealing people's livelyhood through taxes, and shows as heroes the group that takes down said state and returns wealth to its actual owners, the people who worked to produce it... while we commonly think of it as a story that demonizes the rich, who somehow have all this money that shouldn't belong to them, and shows as heroes the group that "steals from the rich and gives to the poor," because the poor deserve that money by virtue of being poor and needing it more.
How the hell did that happen???

You are not off topic at all. Tragedy of the Commons is as much a fallacy as The Lorax, Robin Hood, or even applying thermodynamic laws inappropriately when it comes to complex and open systems. Trying to solve a problem that doesn't exist doesn't mean we're redefining the problem (i.e. demonizing the state because the Sheriff < Richard), I am simply saying that applying a Tragedy of the Commons argument to an artificial structure like money is silly. Tragedy of the Commons can only apply to natural systems like Lotka–Volterra equations. Humans are resourceful enough to find technological solutions to resource issues, not that they always do, but that's another issue entirely NOT related to the resources themselves.

Like wanting to know what the wind chill factor is on steel when the temperature is +19F.

~Bruno~

[Rassah]

Just sell as many patented units on the open market until it is just about saturated at a price equal to say four times your marginal cost. I think you could pawn off several million dollars worth. Now make another, larger batch of units and use these to seize control of 51% of the network, expropriate your customer's mining rights, and earn ~35k USD of revenue a day.

I'm glad THAT scenario is almost impossible 

[DeepBit]

this is hard to imagine working as most of the time it will be more profitable for pool x to also include all the low fees.

That is the point people who naively believe a market will flourish fail to understand.

99.9999999999999999999999999999999999999999999999999999999999999999999999999999 99999999999999999999999999999999999999999999999999999999999999999999999% of the work in solving a block is .... solving the block.  0 transactions or 20,000 transactions the amount of work (and thus cost) is roughly the same.

Actually there is some difference for the pool. Really.

Now this assumes min fee is 0.01 but in actuality it is 1 satoshi.  1 satoshi will guarantee relatively fast confirmations  0 satoshis will guarantee unbelievably slow confirmations and exponentially higher fees will only marginally decrease confirmation time.  Fees will avg 1 satoshi per confirmation until network gets small enough that a monopolist can gain 51% of network, exclude other blocks and raise fee prices creating a cartel.

50 000 satoshi, not 1.

[phelix]

this is hard to imagine working as most of the time it will be more profitable for pool x to also include all the low fees.

That is the point people who naively believe a market will flourish fail to understand.

99.9999999999999999999999999999999999999999999999999999999999999999999999999999 99999999999999999999999999999999999999999999999999999999999999999999999% of the work in solving a block is .... solving the block.  0 transactions or 20,000 transactions the amount of work (and thus cost) is roughly the same.

Actually there is some difference for the pool. Really.

please elaborate

Now this assumes min fee is 0.01 but in actuality it is 1 satoshi.  1 satoshi will guarantee relatively fast confirmations  0 satoshis will guarantee unbelievably slow confirmations and exponentially higher fees will only marginally decrease confirmation time.  Fees will avg 1 satoshi per confirmation until network gets small enough that a monopolist can gain 51% of network, exclude other blocks and raise fee prices creating a cartel.

50 000 satoshi, not 1.

he aint talking about what is hard coded into the standard client. network does not currently enforce a minimum fee.

[Dusty]

[...]
Now this assumes min fee is 0.01 but in actuality it is 1 satoshi.  1 satoshi will guarantee relatively fast confirmations  0 satoshis will guarantee unbelievably slow confirmations and exponentially higher fees will only marginally decrease confirmation time.  Fees will avg 1 satoshi per confirmation until network gets small enough that a monopolist can gain 51% of network, exclude other blocks and raise fee prices creating a cartel.

How to fix it?

Does it need to be fixed?
The (wannabe) monopolist raising fees makes mining more profitable --> new miners try to enter the business taking away the 51% to the (ex) monopolist --> fees get lowered again due to the new competition, and that delicate equilibrium is found again.

[phelix]

As it turns out transactions actually come at a cost:
https://gist.github.com/gavinandresen/5044482  "Back-of-the-envelope calculations for marginal cost of transactions"


So I am not sure if it is really needed any more but I would still like to throw this idea out:

There might be a way to balance a single miners short time interest with all miners long time interest to hold up transaction fees. It allows for something like a limited miners union.

1.) In each block miners vote for a minimum tx fee by the lowest tx fee included. This txFeeMin is averaged over 50 blocks or more.
2.) Only txs with fees above 0.5 * txFeeMinAvg are valid.


The averaging could be done via a moving average (or even together with difficulty changes).

Of course the factor does not have to be 0.5. The averaging time could be 2016 blocks.

Probably it would be more robust to use a median instead of average or an upper limit to prevent artificial increase of txFeeMinAvg by rogue miners.

edited: up to, clarity & brevity, median

[DeathAndTaxes]

0.5 * min?  Wouldn't that force the fee lower and lower until it reached zero.

i.e. avg min fee is 0.001. 0.001 * 0.5 = 0.0005.  So only fees below 0.0005 are valid?

[franky1]

i see in the future when the block chain is too big to manage and all the other blah that people panic about passed their lifetime. people would just leave their coins in a bitcoin address fully verified and then they would dispurse out physical notes or metal coins backed by the verified bitcoin.

thus not requiring the need for miners.

that is one low tech slution.

there are MANY solutions to the problems. so dont worry about 40 years time. just think about the next few years and base things off a "5 year plan" where you change strategies or investment idea's as things change.

[phelix]

0.5 * min?  Wouldn't that force the fee lower and lower until it reached zero.

i.e. avg min fee is 0.001. 0.001 * 0.5 = 0.0005.  So only fees below 0.0005 are valid?

    I meant above. Thanks for pointing that out. What a bummer.


Also I got a new variant:

1.) In each block miners vote for a minimum tx fee by the lowest tx fee included. This txFeeMin is averaged over 50 blocks or more (2016?).
2.) Only 10% of the txs in a block are allowed to be below txFeeMinAvg.

My gut says this might go into the right direction to balance the miner's short term and long term interest.

It would be nice to have some kind of simulation or a small game with real people to quickly test the outcome of such fee rules. And the final test should be with an altcoin.

edit: Percentage of txs in a block might not be such a good idea because miners can simply create txs to themselves (depending on orphan rate).

[grau]

Miner actually vote with every block, since transactions they include define what they expect.

Clients can estimate average fee/byte needed to get into the block by analyzing recent blocks, then using a multiple of that depending on their urgency.

[phelix]

Miner actually vote with every block, since transactions they include define what they expect.

[...]

But currently there is no incentive for a single miner not to include a transaction (besides the increased orphan rates for large blocks) - that's the whole point of this thread.

With a system like this a miner knows that if he does not take a low fee tx the next blocks are also limited in taking low fee tx.

As a miner might see it: punishing low fee tx actually has an effect because it might take a while until it goes into a block.

[whitenight639]

Bitcoin & Tragedy of the Commons

This is perhaps the biggest risk to Bitcoin's future. Does anyone have a convincing argument why this won't happen, or proposals on how to combat it?

Don't fix it, Maybe in 40 years time (with a little blockchain pruning) the actual computation of transaction hashes will be rather trivial and people will run full nodes and mining software on there hand held devices? maybe the small amount of resources needed make it possible to build mining back into the clients, maybe because of the triviallity of required resources, anybody that has interests in bitcoin will run these implementations to keep the network secure even if not for mining, people are altruistic, look at the Fold@home gene mapping cancer research that people run on their computers.


OR

Maybe competing crypto-currencies will be on par with Bitcoins market share, maybe we will have multiple local cryptocurrencies and if Bitcoin dies people will just move, 40 years is a very long time in computing & technology.