Proposal: Idea for a much more stable bitcoin

[naturallaw]

He is saying that, should you mistake the self bootstrapping of a new currency for a speculative bubble (or vice versa) you are likely to make a serious error in investment choices.

Ehh.. I would argue they're one in the same. Either way you're speculating. Could pay off, might not.

[marcus_of_augustus]

Bitcoin is getting more and more stable all the time ... just another non-problem.

Wider adoption will introduce more stability will introduce wider adoption will ....

(see chodpaba time series analysis, the waves are getting broader and longer with each cycle => percentage-wise the volatility will decrease)

As I type, MtGox, Low: 13.51  High: 16.5

That's a change of 22.1% IN ONE DAY.  This is stability?  For perspective, the Dollar moved against the Euro by 0.8% today.  The dollar has fallen against the Euro by about the same 22% margin.... it just took EIGHT YEARS rather than a day.

"more stability" implies relative stability, no? 40% movements have been recorded not so long ago.

For some real perspective, the dollar is 200 hundred years old, bitcoin is 2 years old.

If the dollar was moving 22% in a day against anything after 200 years it would be time to get the hell out.

[marcus_of_augustus]

He is saying that, should you mistake the self bootstrapping of a new currency for a speculative bubble (or vice versa) you are likely to make a serious error in investment choices.

Ehh.. I would argue they're one in the same. Either way you're speculating. Could pay off, might not.

wrong, "speculative bubble" implies a bursting ...

a boot-strapping currency is a bubble that doesn't burst (on a human economic finite time scale)

[CurbsideProphet]

Bitcoin is getting more and more stable all the time ... just another non-problem.

Wider adoption will introduce more stability will introduce wider adoption will ....

(see chodpaba time series analysis, the waves are getting broader and longer with each cycle => percentage-wise the volatility will decrease)

As I type, MtGox, Low: 13.51  High: 16.5

That's a change of 22.1% IN ONE DAY.  This is stability?  For perspective, the Dollar moved against the Euro by 0.8% today.  The dollar has fallen against the Euro by about the same 22% margin.... it just took EIGHT YEARS rather than a day.

"more stability" implies relative stability, no? 40% movements have been recorded not so long ago.

For some real perspective, the dollar is 200 hundred years old, bitcoin is 2 years old.

If the dollar was moving 22% in a day against anything after 200 years it would be time to get the hell out.

No, relative stability doesn't work.  By that logic I could declare I'm starting to dig the next grand canyon.  I dug a 10ft hole in my backyard in 1 day.  Grand canyon took thousands of years of erosion to create.  Relatively speaking, I'm on good pace!

[hugolp]

As I said bitcoin days destroyed would have to be a part of the function. So that transactions of money that just been transacted recently would weigh a lot less. Only when the last transaction of the money was older than some barrier would the transaction weigh in 100% on the total velocity. The time barrier could be set to something like 3 hours, but it would be better to set it dynamically based on the last total velocity. I know the number would not be completely precise, but I think it would still be more so than the velocity when calculated in the real world.

Also a single dude cheating would mean less and less, as more and more people get a hold of money.

Sorry, but this is very very poor. There is no way this is tracking the velocity of money. You are missing all the microtransactions that could be real, and suddenly a guy moves his saved bitcoins (lots of them) to another address and your function would give you a big velocity of money when the guy could be sending it to himself. You are not tracking the velocity of money. The pseudo-anonimity of Bitcoin makes it impossible to track.

[marcus_of_augustus]

Bitcoin is getting more and more stable all the time ... just another non-problem.

Wider adoption will introduce more stability will introduce wider adoption will ....

(see chodpaba time series analysis, the waves are getting broader and longer with each cycle => percentage-wise the volatility will decrease)

As I type, MtGox, Low: 13.51  High: 16.5

That's a change of 22.1% IN ONE DAY.  This is stability?  For perspective, the Dollar moved against the Euro by 0.8% today.  The dollar has fallen against the Euro by about the same 22% margin.... it just took EIGHT YEARS rather than a day.

"more stability" implies relative stability, no? 40% movements have been recorded not so long ago.

For some real perspective, the dollar is 200 hundred years old, bitcoin is 2 years old.

If the dollar was moving 22% in a day against anything after 200 years it would be time to get the hell out.

No, relative stability doesn't work.  By that logic I could declare I'm starting to dig the next grand canyon.  I dug a 10ft hole in my backyard in 1 day.  Grand canyon took thousands of years of erosion to create.  Relatively speaking, I'm on good pace!

Hint:  when you are in a hole stop digging ... you appear to be out of your depth here.

[CurbsideProphet]

Hint:  when you are in a hole stop digging ... you appear to be out of your depth here.

Witty 

When you stated, "more and more stable all the time," it implies some sort of linear progression towards stability.  At Bitcoin's current volatility, I don't see how you can come to that conclusion.  Just my opinion. 

[Babylon]

This would be a change away from one of the basic concepts bitcoin is founded on.  However you could take the sourcecode and make your own velocitycoin, or whatever, that works in this way.

[LeonGeeste]

Just saw this proposal today.  My reaction: NO, NO, NO, NO, NO.

Even if everything you say about your idea is true, that doesn't change the fact that it would be, as you seem to accept, changing the rules.  It would be breaking the promise made to every single user that there would be no money creation beyond that specified by the protocol.  Not acceptable.  So, at most, you've justified starting a *different* bitcoin-like currency, and I wish you the best of luck in getting such an inflationary currency off the ground, though that will be kinda difficult.

But please, only impose this inflation on people who have accepted it; don't change horses midstream on a currency that only has its current popularity because of a promise that it would pull stunts like yours.

Setting all that aside: as others have mentioned, such a system can be gamed through fake transactions, and your time-money system does not prevent this, it just delays it, and encourages the earliest-possible overloading of the block-chain with pointless transactions.

In your defense, this is not an error on your part, but an error of all the economists who advocate NGDP targeting, and its shortcomings here are a direct implication of the unappreciated shortcomigns of NGDP targeting in regular money.

If you want to stabilize the value of Bitcoin, there's only one way: widespread agreement regarding exactly how useful Bitcoin will be.  And to do that, you need to remove uncertainty.  The protocol itself does a lot already to accomplish this: by having a predictable money supply (appropriately defined).  Introducing a new source of uncertainty -- how many more will be produced by your new scheme -- only increases volatility.

[Stuffe]

Bitcoin is getting more and more stable all the time ... just another non-problem.

Wider adoption will introduce more stability will introduce wider adoption will ....

(see chodpaba time series analysis, the waves are getting broader and longer with each cycle => percentage-wise the volatility will decrease)

Before central banks most currencies in Europe were bound to gold or silver. The currencies were mainly used for trade (as opposed to speculation), like you foresee the bitcoin in the future. But the currencies were certainly not stable. I recently read about the influx of gold when the america was discovered and how the price of gold fell to 1/3 of its prior value. Extreme case, but I think you get the point. Bitcoin will get more stable, but will never really be stable.

What I suggest is to change the rules a bit. So instead of just having the miners create 50BTC ever 10 minutes, we set the creation rate based on the velocity of the money.

This is already in the bitcoin protocol. Over time, the block reward is lowered. Transaction fees follow velocity. So just wait, everything will be OK.

Really good point, it will help to stabilize, but surely it wont be enough. Do you really think so and why?

Such a proposal would be subject to manipulation, which disqualifies it for consideration.  However, if you insist that your idea is better, start your own blockchain and change the name of your currency.  We will let the market decide.

There are ways which you could make this more or less manipulation immune.

Bitcoin is getting more and more stable all the time ... just another non-problem.

Wider adoption will introduce more stability will introduce wider adoption will ....

(see chodpaba time series analysis, the waves are getting broader and longer with each cycle => percentage-wise the volatility will decrease)

As I type, MtGox, Low: 13.51  High: 16.5

That's a change of 22.1% IN ONE DAY.  This is stability?  For perspective, the Dollar moved against the Euro by 0.8% today.  The dollar has fallen against the Euro by about the same 22% margin.... it just took EIGHT YEARS rather than a day.

"more stability" implies relative stability, no? 40% movements have been recorded not so long ago.

For some real perspective, the dollar is 200 hundred years old, bitcoin is 2 years old.

If the dollar was moving 22% in a day against anything after 200 years it would be time to get the hell out.

The dollar is stable because it is more or less based on the system I am proposing.

As I said bitcoin days destroyed would have to be a part of the function. So that transactions of money that just been transacted recently would weigh a lot less. Only when the last transaction of the money was older than some barrier would the transaction weigh in 100% on the total velocity. The time barrier could be set to something like 3 hours, but it would be better to set it dynamically based on the last total velocity. I know the number would not be completely precise, but I think it would still be more so than the velocity when calculated in the real world.

Also a single dude cheating would mean less and less, as more and more people get a hold of money.

Sorry, but this is very very poor. There is no way this is tracking the velocity of money. You are missing all the microtransactions that could be real, and suddenly a guy moves his saved bitcoins (lots of them) to another address and your function would give you a big velocity of money when the guy could be sending it to himself. You are not tracking the velocity of money. The pseudo-anonimity of Bitcoin makes it impossible to track.

.

While not precise, the velocity would be calculated far more precise than that which is used to control the inflation of say, the dollar. They calculate it using a consumer price index, which is itself just a survey of goods and services and how they change. Not very precise, but it works pretty ok.

This would be a change away from one of the basic concepts bitcoin is founded on.  However you could take the sourcecode and make your own velocitycoin, or whatever, that works in this way.

Which concept exactly? Instability?

Just saw this proposal today.  My reaction: NO, NO, NO, NO, NO.

Even if everything you say about your idea is true, that doesn't change the fact that it would be, as you seem to accept, changing the rules.  It would be breaking the promise made to every single user that there would be no money creation beyond that specified by the protocol.  Not acceptable.  So, at most, you've justified starting a *different* bitcoin-like currency, and I wish you the best of luck in getting such an inflationary currency off the ground, though that will be kinda difficult.

But please, only impose this inflation on people who have accepted it; don't change horses midstream on a currency that only has its current popularity because of a promise that it would pull stunts like yours.

Setting all that aside: as others have mentioned, such a system can be gamed through fake transactions, and your time-money system does not prevent this, it just delays it, and encourages the earliest-possible overloading of the block-chain with pointless transactions.

In your defense, this is not an error on your part, but an error of all the economists who advocate NGDP targeting, and its shortcomings here are a direct implication of the unappreciated shortcomigns of NGDP targeting in regular money.

If you want to stabilize the value of Bitcoin, there's only one way: widespread agreement regarding exactly how useful Bitcoin will be.  And to do that, you need to remove uncertainty.  The protocol itself does a lot already to accomplish this: by having a predictable money supply (appropriately defined).  Introducing a new source of uncertainty -- how many more will be produced by your new scheme -- only increases volatility.

First of, it will NOT be an inflationary coin, as I wrote, I am suggesting stable deflation. That said, I understand your point that it is changing the rules and some people are not comfortable with that. As for the fake transactions, the time can be adjusted. Still with three hours, you would need to have a very substantial amount of the total amount of bitcoins to be able to make any real impact. Also that impact would be negated the second you stop transacting every 3 hours. Not a real problem.
NGDP targeting? This is a currency, "Bitcoin land" is not really real, thus it has no National Gross Domestic Product. I am actually proposing inflation/deflation targeting.

[Babylon]

build it Stuffe,  the source code is there.  If it is actually better than bitcoin it will replace it.

[Stuffe]

build it Stuffe,  the source code is there.  If it is actually better than bitcoin it will replace it.

Would love to, but unfortunately im not a coder, just an economist.

[cunicula]

Such a proposal would be subject to manipulation, which disqualifies it for consideration.  However, if you insist that your idea is better, start your own blockchain and change the name of your currency.  We will let the market decide.

I don't like aspects of the idea (e.g. the hypothesized voluntarily mining at a loss is absurd), but I hate responses like MoonShadow's one hundred times more.

The bitcoin status quo will have an advantage due to network effects. To say let the market decide is akin to saying let the market decide if microsoft makes the world's best OS. What the market really decides is whether some other OS is so much better than Windows that that it can overcome microsoft's network effect. This is a much much bigger hurdle than simply delivering a better OS. The long-run inefficiency associated with these hurdles is one of the main reasons we have antitrust laws in place.

[markm]

build it Stuffe,  the source code is there.  If it is actually better than bitcoin it will replace it.

Would love to, but unfortunately im not a coder, just an economist.

Not necessarily an insurmountable problem. What exactly is the formula to be used to determine the number of coins a specified block number should mint? This must of course be a figure that can be computed by a client that is downloading the blockchain for the first time, checking each block as it goes to make sure the cirrect number of coins were minted in that block.

How many servers can you deploy to listen on a port for people running a client for your new currency? Or what bounty will be py people to set these up for you? (Unthinkingbit offered one bitcoin each for five miners to do so, and has not yet five miners doing so I don't think, so you might like to offer a little more than that?)

Once we have at least a small network running your new coin, I can also add it to the repertoire of my trading bots, the more blockchain-based currencies they are able to deal in the better.

-MarkM-

[cunicula]

The objection to the transaction tax is that it would decrease the value of bitcoin and could thus augment rather than counter inflationary pressure.. This is easy to solve. Keep the tax rate on transactions constant., but change tax revenue distribution rules. If there is inflation destroy the tax revenue. If there is deflation distribute the tax revenue to miners.

[Stuffe]

Such a proposal would be subject to manipulation, which disqualifies it for consideration.  However, if you insist that your idea is better, start your own blockchain and change the name of your currency.  We will let the market decide.

I don't like aspects of the idea (e.g. the hypothesized voluntarily mining at a loss is absurd), but I hate responses like MoonShadow's one hundred times more.

The bitcoin status quo will have an advantage due to network effects. To say let the market decide is akin to saying let the market decide if microsoft makes the world's best OS. What the market really decides is whether some other OS is so much better than Windows that that it can overcome microsoft's network effect. This is a much much bigger hurdle than simply delivering a better OS. The long-run inefficiency associated with these hurdles is one of the main reasons we have antitrust laws in place.

Thank you cunicula, I too very much dislike those counterproductive comments.

But I am not suggesting that anyone would be mining at a direct loss ever (no one would ever have their bitcoins destroyed for mining.) Of course the loss will be real in that electricity costs money and miners might just turn off their rigs to save energy if transaction fees are not high enough. That is indeed a problem, but that same problem will also arise with the current bitcoin when no more coins can be mined. Actually that sounds like a real danger for today's bitcoin, since a lot of rigs will be turned off, leaving the network so much more vulnerable to the "51% attack".
As for a solution (under my system) maybe miners could be promised some of their returns to be delivered in the future, at what ever the rate is at that time? Like, you will get 25% of what the rate is today right now, then you will get 25% of what the rate is next month, at next months rate, then you will get 25% in two months at that times rate and finally the last 25% in three months at whatever the rate is at that time. It will take some time to get all the coins (length of time can of course be adjusted and should be decided by the miners themselves).

I think you raised a valid concern there, if you see other problems or things you dislike, please post them

build it Stuffe,  the source code is there.  If it is actually better than bitcoin it will replace it.

Would love to, but unfortunately im not a coder, just an economist.

Not necessarily an insurmountable problem. What exactly is the formula to be used to determine the number of coins a specified block number should mint? This must of course be a figure that can be computed by a client that is downloading the blockchain for the first time, checking each block as it goes to make sure the cirrect number of coins were minted in that block.

How many servers can you deploy to listen on a port for people running a client for your new currency? Or what bounty will be py people to set these up for you? (Unthinkingbit offered one bitcoin each for five miners to do so, and has not yet five miners doing so I don't think, so you might like to offer a little more than that?)

Once we have at least a small network running your new coin, I can also add it to the repertoire of my trading bots, the more blockchain-based currencies they are able to deal in the better.

-MarkM-

Are you saying you would help me build it? I would be willing to write a very thorough specification if yes

The objection to the transaction tax is that it would decrease the value of bitcoin and could thus augment rather than counter inflationary pressure.. This is easy to solve. Keep the tax rate on transactions constant., but change tax revenue distribution rules. If there is inflation destroy the tax revenue. If there is deflation distribute the tax revenue to miners.

Another good point. I don't really like the idea of a forced tax though, it is hard to say if it would have enough impact, but the transaction fees could be taxed. And as inflation causes more and larger transactions, the amount of total transaction fees would therefore also increase. Still as transaction fees are voluntary that means people can always transact for free if they want to. What do you think?

[cunicula]

I think the tax is highly desirable. As with bitcoin why would anyone want to accept payment in a currency that has a substantial risk of inflating. Most people are not speculators by nature. The protection from inflation offered by a tax is worthwhile. Without a tax, pegging generation to velocity would reduce volatility, but it could not stop inflation due to a fall in demand.

A few points:
1 ) voluntary transaction fees make no sense to me. Transactions add information to the blockchain permanently. They impose costs on everyone using the sytem (storage costs, bandwidth costs). This should not be free.
2) a big problem is measurement of velocity. A side benefit of transaction fees is that they will make velocity measures more reliable.
3) Currently miners are paid through seignorage (issuing coins), this is equivalent to a tax on the rest of the user base. Replacing a part of the seignorag tax with a transaction fee tax would not make users worse off in general. (the incidence of the two taxes is different but they are both taxes. more on this later.)
3) Miners will always need to be rewarded. Therefore any velocity based system could never drop the coin generation rate to zero.
4)the ability of the transaction tax to prop up value depends on being able to destroy more coins than you create each period. If you can consistently destroy more coins than you create, you would be able to support the price of a currency unit even if demand for the currency falls. Of course if demand falls to zero then you are shit out of luck.
5 ) pegging generation to velocity would enable such a system. The tax revenue per period is the product of velocity and the tax rate. As long as currency generation is less than tax revenue the system can destroy more coins than it creates each period.
6) the tax rate would have to be very low to keep the currency useful and prevent evasion (e.g. Sending wallet files as email attatchments). I would suggest a 0.1% levy on all sends. There should be a very small fee per kb as well which would only apply to microtransactions.
7) currency generation would be smaller than potential destruction (currency generation <  velocity *tax rate).
Miners would earn most of their income from the tax. In the case of inflation, mining income would fall by more than half because the tax would be devoted to coin destruction. In the case of deflation, the tax would be used to pay miners.
9) coin creation and destruction would continue indefinately.
10) the tax incidence of a transaction tax and seignorage are different. Seignorage is born by people who hoard their coins. Transaction taxes are born by people who spend coins.
11) rolling out a coin generation mechanism based on velocity is tricky. One option is to set a minimum coin generation rate. This would be set extremely low, so that hopefully it would only be binding in the currencies initial month or so.

[Stuffe]

1 ) voluntary transaction fees make no sense to me. Transactions add information to the blockchain permanently. They impose costs on everyone using the sytem (storage costs, bandwidth costs). This should not be free.

There is a voluntary transaction fee being used at the moment. The higher the fee you pay, the faster miners will serve your transfer.

2) a big problem is measurement of velocity. A side benefit of transaction fees is that they will make velocity measures more reliable.

Velocity should not be hard at all to calculate, its all in the block chain. If you are talking about bogus transfers, then your point is correct, but I really don't think those are that big of a problem (and the voluntary transaction fee was introduced exactly to fight this in the first place).

3) Currently miners are paid through seignorage (issuing coins), this is equivalent to a tax on the rest of the user base. Replacing a part of the seignorag tax with a transaction fee tax would not make users worse off in general. (the incidence of the two taxes is different but they are both taxes. more on this later.)

Well, it sounds really bad to "tax" and if this was ever to be adopted, it would need to be accepted by a majority of anarchists and liberal types on this board and they don't generally know that mining is the same as inflating, which is the same as tax. Also people would have a slightly lower incentive to transact, so it wouldn't have exactly the same effect, but yea, almost.

3) Miners will always need to be rewarded. Therefore any velocity based system could never drop the coin generation rate to zero.

Did you see my comment on spreading out the mining profit over time? The idea is, profit now at current rate and a promise of profit at later dates at future rates. Regardless of whether the rate is 0 now, the promise of return on a later date will still be worth something. If you did see this proposal but don't like it, please elaborate.

I expect to see a lot huge increase in the velocity over the next couple of years so I don't think controlling inflation will be a big problem long term. Also it is generally excepted that the coin is deflationary by nature. Short term inflation (sudden price drops) will quickly be eliminated by speculators (especially since they have the source code and can very accurately calculate the probable mining rate at the next period, in real time).

Another tool to further be able to fight inflation would be to calculate a "deflation margin" based on the volatility of the velocity. This just means if the money velocity was volatile last period (prices were volatile), we would allow for more deflation in this period. This would be to ensure that we don't pump out a lot of money this period to hit our deflation target, just to find out that next period the coin is very inflationary and we wish we could take all those coins back we just pumped out.

Ultimately, whether or not your idea is good comes down to whether or not taxing the already existing transaction fees and the "deflation margin" will be enough to cope with inflation or not. And to be honest I am not sure that it is.

[cunicula]

Actually I love the idea of delayed rewards for miners. Of course these are also inflationary, they  just add less to the money supply than their face value. If you search my posting history (click on my name) you will see that i am the official forum fanatic on this topic. Delayed rewards have many applications above and beyond managing money supply growth. I think control over the money supply will be much much weaker if the system cant destroy coins.

As far as libertarians and their views go, i very much doubt that they are the future cryptocurrency user base. They dominate the forum, but if cryptocurrency succeeds they will be a tiny minority of the user base. Much better IMO to appeal to finance types and regular merchants. These are the people who can make decisions that cause cryptocurrency to take off.

However, I dont have to get my way 100% to get behind a project. Give me a system with tradeable claims to delayed rewards and I'll become a fanatic proponent.

[BTConomist]

Of course the loss will be real in that electricity costs money and miners might just turn off their rigs to save energy if transaction fees are not high enough. That is indeed a problem, but that same problem will also arise with the current bitcoin when no more coins can be mined. Actually that sounds like a real danger for today's bitcoin, since a lot of rigs will be turned off, leaving the network so much more vulnerable to the "51% attack".

This is why bitcoin should never be confused with an investment asset, commodity or even a collector's item. It's only a currency (a medium of exchange, or an accounting ledger, if you will). It can only exist if the number/volume of transactions is high enough to compensate miners for running their rigs. Hoarding bitcoin (either by miners or users) would only lead to its demise.

Measuring bitcoin in how many (or how much) fiat currencies it can buy is as useless as measuring it in how many bananas it can buy: at the end, it's all about whether the supply of those goods (services) is increasing or decreasing with time.

As for a solution (under my system) maybe miners could be promised some of their returns to be delivered in the future, at what ever the rate is at that time? Like, you will get 25% of what the rate is today right now, then you will get 25% of what the rate is next month, at next months rate, then you will get 25% in two months at that times rate and finally the last 25% in three months at whatever the rate is at that time. It will take some time to get all the coins (length of time can of course be adjusted and should be decided by the miners themselves).

Miners should realize they hold the key to bitcoin's success.
They are the central banks of bitcoin economy.
It's time they start acting like it.

Update:
For more, see http://forum.bitcoin.org/index.php?topic=29565.msg373207#msg373207