A proposed solution to adjust for lost Bitcoins: wallet 'heartbeats'

[BitterTea]

Hey guys, I have an idea!

You should collaborate on a new block chain using these rules. I seriously doubt you're going to convince enough people that this belongs in Bitcoin, but you can start a competing crypto currency with whatever changes you want. Bitcoin is open source, go build something!

[ascent]

No, they are not. That is why I showed you the geometric series, which is an infinite series of positive values that add up to a constant, even if you keep adding FOREVER.

Saying that more and more bitcoins get lost does NOT mean that the number or ratio of bitcoins still in circulation goes towards zero! Not at all. That is the point. It might mean that, but in all probability it does not mean that. If fact, as I have pointed out above, the yearly ratio of bitcoins lost is bound to converge toward zero very fast, which implies that the number of bitcoins still in circulation will remain above a large constant, say 10 million, FOREVER. That is why your problem will never be a problem.

I never stated that a constant or near constant number of Bitcoins would be lost per year or per unit time. What I said was this:

M = total number of Bitcoins ever minted
C = total number of Bitcoins not lost
L = total number of Bitcoins truly lost
t = time

Per any unit of time you wish to use, L at t + 1 will be >= L at t.

Logically, then, L will get bigger over time. From that, it follows L will eventually approach M.

Let's assume that you agree that a certain percentage of unlost Bitcoins will be lost every year. Is that not an unreasonable assumption? No matter what number you choose, if you continue with the following calculation, the end result is the same:

C at t = M
C at t + 1 = (C at t) * x where x = some constant number less than 1.

[riush]

how about some service that does the heartbeating for you? someone might still find the login info to that on grandpas old laptop...

[BitterTea]

I never stated that a constant or near constant number of Bitcoins would be lost per year or per unit time. What I said was this:

M = total number of Bitcoins ever minted
C = total number of Bitcoins not lost
L = total number of Bitcoins truly lost
t = time

Per any unit of time you wish to use, L at t + 1 will be >= L at t.

Logically, then, L will get bigger over time. From that, it follows L will eventually approach M.

Let's assume that you agree that a certain percentage of unlost Bitcoins will be lost every year. Is that not an unreasonable assumption? No matter what number you choose, if you continue with the following calculation, the end result is the same:

C at t = M
C at t + 1 = (C at t) * x where x = some constant number less than 1.

Bzzzzt. You fail to factor in the value of a bitcoin. As the value increases, the rate at which they are lost will decrease, as people will take additional measures to secure them. If R_L approaches 0, L will approach some value less than M.

[tubro]

I never stated that a constant or near constant number of Bitcoins would be lost per year or per unit time. What I said was this:

M = total number of Bitcoins ever minted
C = total number of Bitcoins not lost
L = total number of Bitcoins truly lost
t = time

Per any unit of time you wish to use, L at t + 1 will be >= L at t.

Logically, then, L will get bigger over time. From that, it follows L will eventually approach M.

No, this is not logical. This is an error in your thinking. Just because something becomes less and less, doesn't mean it will eventually become zero. This is a little tricky to grasp for the non-mathematician, but is nevertheless true and important.

[AntiVigilante]

As we all know, the true number of Bitcoins in circulation will always be some unknown number less than the total number of Bitcoins mined. This is not ideal. Wallets can be lost, deleted, or the underlying media on which they are stored on can be damaged beyond recovery.

Are Bitcoins in it for the long haul? Or are they just a five year experiment?

I'm proposing a solution: wallet 'heartbeats'. I'm not sure if it is technically compatible with the existing software infrastructure, but I think it might be. If we define a heartbeat as connecting to the network, then perhaps wallets should give a heartbeat at least every seven years in order to remain valid. Any coins in a wallet that has not connected to the network for seven years become invalid and are made available to be remined, by some method that allows for their easy mining. Because I don't fully understand the mechanics behind the software, I don't know if this is possible, but if it is, I believe the idea has merit.

You could have symbolic links between addresses and unix-like permissions.

As long as the symbolic link exists the BTC would be transferable remotely from address to address without the original wallet. Bye bye lost coins.
Unix permissions would also prevent the theft that occurred at MtGox by allowing you to block transactions not authorized by you. You would pre-generate transactions which would be filled in by MtGox. That is you would control MtGox from the client. MtGox would not be able to create transactions without your wallet being online and sending requests.

Problem, criminal?

I think I'll work on this. After JSON is fully the format of the system. At this point I can't navigate the code like others can.
Unix permissions: Receive, Send, Operate

[ascent]

No, this is not logical. This is an error in your thinking. Just because something becomes less and less, doesn't mean it will eventually become zero. This is a little tricky to grasp for the non-mathematician, but is nevertheless true and important.

Oh, I see now! I'm not a mathematician, therefore, I can't grasp the concept fully. First of all, it does not need to become zero for the problem to exist. Secondly, the key point is "approach zero". I shall now use the same intonation as your post above: I realize this is tricky for you to grasp, therefore, I suggest that instead of trying to sideline the argument into whether the supply will reach zero or some quantity that is not exactly zero, which is a wrong assumption on your part anyway, that you reconsider the real premise put forth in this thread, and find a means to object to it, instead.  That is, if it's not too tricky for you.

[cschmitz]

As we all know, the true number of Bitcoins in circulation will always be some unknown number less than the total number of Bitcoins mined. This is not ideal. Wallets can be lost, deleted, or the underlying media on which they are stored on can be damaged beyond recovery.

Are Bitcoins in it for the long haul? Or are they just a five year experiment?

You assume that bitcoin has a problem with availability or granularity. Both is horribly false in the forseeable long term future.
The price of a bitcoin unit of a certain deciaml within the p2p accounting system is set via supply and demand and zero via regulation, which is a key selling point of bitcoin.
You are trying to add an automated regulation mechanism to solve a problem that doesnt exist, a very uncompelling case.

That is, if it's not too tricky for you.

You have been around for a week, maybe some humility would help?

[ascent]

You have been around for a week, maybe some humility would help?

Not when people like you come in and start talking about divisibility. Read the thread. It has been stated numerous times that divisibility has nothing to do with it.

[ascent]

You assume that bitcoin has a problem with availability or granularity.

It's really annoying to have someone say that I assume something, when I have clearly stated the opposite. You only undermine your own reputation, and undermine it further by claiming the importance of tenure. And you wonder why I don't care how long someone has been around. Content matters. Don't hide behind tenure. I would accord equal respect to any member based upon what they are saying, not how long they have been around.

[cschmitz]

You have been around for a week, maybe some humility would help?

Not when people like you come in and start talking about divisibility. Read the thread. It has been stated numerous times that divisibility has nothing to do with it.

No, the root of the problem that you are trying to find and cure would be divisibility. Your initial posting is full of "i dont know" and in that regard very valid. You fail to actually identify a problem yet want to make a connection mandatory, however long the duration might be. Bitcoins will be valued at supply and demand related rates. This is possible as long as divisibility is sufficient, hence the problem you are trying to cure is not there. You try to solve a problem that doesnt exist by introducing a system that gives alot of people with offline savings a maintenance overhead that is neither desireable workload wise nor smart in the pseudo anonymous context. But of course, you wouldnt know about that.

If in the very long term bitcoin is still around, the system can be branched to adress that issue without changing the rules, if 50+% of the network agrees to it. You would know that if you would take time to understand a system before suggesting LOOONG term solutions for LOOONG term issues that may or may not be there.

[bji]

No, the root of the problem that you are trying to find and cure would be divisibility. Your initial posting is full of "i dont know" and in that regard very valid. You fail to actually identify a problem yet want to make a connection mandatory, however long the duration might be. Bitcoins will be valued at supply and demand related rates. This is possible as long as divisibility is sufficient, hence the problem you are trying to cure is not there. You try to solve a problem that doesnt exist by introducing a system that gives alot of people with offline savings a maintenance overhead that is neither desireable workload wise nor smart in the pseudo anonymous context. But of course, you wouldnt know about that.

How will the market value bitcoins when the total number in circulation is unknown?  Does it matter how deeply the bitcoins can be divided if no one knows if there are 10 million bitcoins in circulation or 20 million?

This is an honest question.  Maybe someone with a better grasp of economics can tell me if it matters at all whether the total number of units of a currency that exists is known or not known; maybe it doesn't matter because people value the currency based on its current supply, by which I mean the availability of units of currency to purchase, not on the total number in circulation.

Like I said before I personally would find it satisfying to have a system where no units of currency could ever be lost and with Bitcoin we could have that; but practically speaking the difference between 0.5% deflation and 0.0% deflation is probably not very important on anything other than millenia time scales.

[ascent]

No, the root of the problem that you are trying to find and cure would be divisibility.

It's amazing how many people insist I think divisibility is an issue.

[BitterTea]

No, the root of the problem that you are trying to find and cure would be divisibility.

It's amazing how many people insist I think divisibility is an issue.

I don't. You still haven't addressed my point that as the value of a bitcoin increases, the incentive to protect it from loss (or theft, but that doesn't decrease the bitcoin in circulation) increases. This, I think, will lead to the slowing of loss over time.

[AntiVigilante]

No, the root of the problem that you are trying to find and cure would be divisibility.

It's amazing how many people insist I think divisibility is an issue.

I don't. You still haven't addressed my point that as the value of a bitcoin increases, the incentive to protect it from loss (or theft, but that doesn't decrease the bitcoin in circulation) increases. This, I think, will lead to the slowing of loss over time.

Right now that incentive is being expressed by feeding this thread not protecting better.

[bji]

No, the root of the problem that you are trying to find and cure would be divisibility.

It's amazing how many people insist I think divisibility is an issue.

I don't. You still haven't addressed my point that as the value of a bitcoin increases, the incentive to protect it from loss (or theft, but that doesn't decrease the bitcoin in circulation) increases. This, I think, will lead to the slowing of loss over time.

I am not so sure.  If bitcoin becomes widespread in use, then I would think that there would be a spreading around of bitcoins more than anything else.  So if right now there are 10000 people with an average of 200 bitcoins each (just to pick some numbers), then if bitcoin becomes very popular then there might be 1000000 people with on average 2 bitcoins each.  Now there are many, many more opportunities for people to lose bitcoins.  We can see that even today with a high value of ~$20 per bitcoin, some people lose them due to carelessness (the recent heist because a person didn't bother to encrypt their wallet file comes to mind), and even web sites that were pretty much printing money (mtgox) don't do a good job of securing their coins and (possibly - hopefully!) killed their goose that was laying golden eggs.

So clearly just because bitcoins are valuable, it doesn't mean that there won't be careless people who lose them.  And if you spread the bitcoins around, there will be less value to lose per individual 'wallet', but many more wallets.  I personally think it's kind of a wash, and I would expect a constant rate of loss that actually exceeds that of paper money since digital currency like bitcoins is so much easier to destroy (with a single keystroke you can destroy a large quantity of bitcoins instantly; with paper money you'd probably have to throw a big suitcase into the ocean or something, which takes alot more effort).

[bji]

No, the root of the problem that you are trying to find and cure would be divisibility.

It's amazing how many people insist I think divisibility is an issue.

I don't. You still haven't addressed my point that as the value of a bitcoin increases, the incentive to protect it from loss (or theft, but that doesn't decrease the bitcoin in circulation) increases. This, I think, will lead to the slowing of loss over time.

Right now that incentive is being expressed by feeding this thread not protecting better.

The end result of this thread could be better protection if people like you would just go away and leave the discussion to people willing to discuss new ideas without resulting to trivialization and demeaning them.

[Maged]

I agree with you that it is a weakness of the system that there are no ways to distinguish whether a coin is loss or simply hoarded. From the economic view, it may raise concern of sudden discovery of the old wallet in the future. However, your solution is not good and it is too complex to be accepted into the current protocol.

My proposal: Unmoved Bitcoin with age more than, say, 40 years are marked as lost and cannot be used for further transaction.

Imposing a time limit has many advantage over the supposed infinite time model:
1. All cryptography keys in use today need to update periodically, for security reason. Leaving the same public key in block for a long time will simply increase the risk of being hacked. So moving the Bitcoin into a new wallet could secure their Bitcoin.
2. It allows the network drop old blocks.
3. This proposal can be implemented very easily with few lines.
4. The proposal implies that value of the lost Bitcoin will be automatically go to all others wallet.
5. It allows the differentiation between lost Bitcoin and hoarded Bitcoin to reduce uncertainty.
6. They should know that Bitcoin can be lost so that they wont put everything into Bitcoin.

There is no such need to reclaim the lost Bitcoin or allow for new mining. Otherwise, if there is 100000 Bitcoin lost for 40 years, you just give the miner with the ability to manipulate the market which make no difference with someone who discovers it. Also, the time interval is should not be too short, say, 1 year. It will only create an instability for the Bitcoin economy. For example, you might want to destroy others wallet in order to maximize your own profit.

Using the proposal, everyone else is actually gaining a very tiny value of their own Bitcoin and it is the most fair method to handle lost of Bitcoin. Remember that the creation of Bitcoin out of nothing is simply a way to make an initial distribution of Bitcoin, there is no need to mining in a sustainable system anymore in the future. I know someone do not like to see that the decimal in their account is smaller and smaller over year. But it is just a game of number, you should really divided by the maximum number of Bitcoin.

If 99% of bitcoin is marked as lost, we can simply multiply 100 to the balance in the bitcoin client. 

I agree that this whole issue is solved by simply marking coins as lost. I also agree that we will have to do this eventually. The great news is, unlike the proposal to reintroduce the coins, this is actually compatible with the current blockchain! For this to work, we don't need to have everyone upgrade, we just need over 50% of the miners to agree not to include any transaction using expired coins into a block and reject any block that does. We could literally implement this policy today. Because of that, I feel that there is no reason to worry about this until it becomes an issue. Whether that issue is advances in cryptanalysis or because more than 10% of the Bitcoin supply is in an unknown state, is irrelevant.

[ascent]

I am not so sure.  If bitcoin becomes widespread in use, then I would think that there would be a spreading around of bitcoins more than anything else.  So if right now there are 10000 people with an average of 200 bitcoins each (just to pick some numbers), then if bitcoin becomes very popular then there might be 1000000 people with on average 2 bitcoins each.  Now there are many, many more opportunities for people to lose bitcoins.  We can see that even today with a high value of ~$20 per bitcoin, some people lose them due to carelessness (the recent heist because a person didn't bother to encrypt their wallet file comes to mind), and even web sites that were pretty much printing money (mtgox) don't do a good job of securing their coins and (possibly - hopefully!) killed their goose that was laying golden eggs.

So clearly just because bitcoins are valuable, it doesn't mean that there won't be careless people who lose them.  And if you spread the bitcoins around, there will be less value to lose per individual 'wallet', but many more wallets.  I personally think it's kind of a wash, and I would expect a constant rate of loss that actually exceeds that of paper money since digital currency like bitcoins is so much easier to destroy (with a single keystroke you can destroy a large quantity of bitcoins instantly; with paper money you'd probably have to throw a big suitcase into the ocean or something, which takes alot more effort).

bji has it one hundred percent correct. If the value of Bitcoins is to rise, it will be due to increased adoption, and more likely, widespread adoption. Remember, the increasing value today is due to speculation that they will become a useful vehicle for commerce, not because they are currently a useful vehicle for commerce.

Widespread adoption will result in increased value, but as bji says, it will not likely result in a decreased rate of loss. In the far future, widespread use might mean hundreds of millions, or billions of users. Regardless of increased value, due to diminishing numbers in circulation, the increased user base will cause any one wallet to likely have a relatively constant average value, based on external economic factors. This wallet value will be no different than average values of checking accounts and wallets in use today. Increased valuations of Bitcoins is unlikely to mean that any one wallet in existence is protected at a security level greater than a wallet today.

The widespread use, if it comes to fruition, will mean exposure all across the world at a high density level, and even the Solar System, if we are talking about the far future. Bitcoins will be subject to all the same accidents that can befall physical objects today, such as natural disasters (how many Bitcoins were lost in the recent Japanese tsunami?), but Bitcoins are also subject to loss via mechanisms that usually aren't an issue with normal physical objects: hardware failure of the hosting device, software viruses, accidental deletion, malicious deletion, and so on.

I contend that the most likely vector for Bitcoins to gain widespread use will be if they supplant the unstable currencies of the Third World, usable on smartphones, allowing near instant commerce between the citizens of the Third World, across state lines. This is the exact scenario which creates a situation ripe for loss, but is also likely the only real hope Bitcoins have for mass adoption.

If Bitcoins fall out of favor, either due to competing currencies, the increasing uncertainty that has been described here, or actually fail to gain widespread use, then the speculation which attaches an increasing value to them today will not guarantee a high valuation for them in the future,.

[tubro]

Oh, I see now! I'm not a mathematician, therefore, I can't grasp the concept fully. First of all, it does not need to become zero for the problem to exist. Secondly, the key point is "approach zero".

It might sound condescending. But it's not. It is really a concept that is hard for non-mathematicians to understand, because people rarely get to think about infinite things.

See, you still haven't understood it: The number of active bitcoins will NOT "approach zero". Not at all. It will approach some large positive number. Ofcourse, it is hard so estimate the exact number today, but I believe it will be more than half the number of total bitcoins, that is more than 10.5 million.