Removing old coin uncertainty

[gmaxwell]

I agree that supply uncertainty is a concern and a mistake in the original system. It isn't clear to me that it can be resolved without causing greater damage, since the inviolability of our promises is even more important— if Bitcoin operates on mere popular whim it offers little more than the official currency of a democratic country.

What concerns me most in this space is the cryptographic break concern:  Lets imagine that at some point 90% of the coins are lost... no biggie, we just trade in smaller units. One small country costs 10 BTC, etc.  And then someone comes up with a way of breaking ECDSA and can suddenly recover hundreds of thousands of long lost coins and introduce them into circulation at will... perhaps many someones.

No cryptographer will claim that our current signature algorithm will be secure forever.  Bitcoin is forward adaptive and can gain new signature algorithims without a hard fork.  Non-lost coins should be migrated to new, more secure, signatures long before it's an issue.  But no one can move the lost coins so their reintroduction into circulation via cryptographic compromise could be uncontrolled and devastating to the Bitcoin economy.

The obvious fix for this is that any crypto upgrade would require coins to be moved after some suitable period... but as you've seen here, people are _very_ hostile to the idea.

I guess my thinking on this is that the idea that a system can be free from human intervention is a bit of an unrealistic fantasy, though one I frequently enjoy. A failure to intervene when doing so would be rational and necessary is also a kind of intervention, and the Bitcoin system may someday die from it. What does it matter if your coins are not "stolen" from you in the broadest possible sense if dogmatic adherence to that principle ultimately results in the coins being worthless?  Time will tell.

[1714713428]

1714713428

[1714713428]

1714713428

[1714713428]

1714713428

[1714713428]

1714713428

[Birdy]

@gmaxwell: another +1, well written
Reflex-screaming "thief!!" isn't a healthy discussion. There may be times when an intervention would be better than doing nothing. But it's not now.

[jratcliff63367]

I am of the opinion that a lot of the uncertainty can be resolved with statistical analysis of the blockchain.

It's a personal project I hope to spend more time on whenever I have some free time.

Here is what I would like to see.

A graph of stale coins that starts at 100% of all outstanding coins in January 1, 2013.  Then, each time coins associated with a particular address has any kind of a transaction then it is considered essentially 'alive' (no longer a zombie coin).  This graph would go down to land wherever we are at today (coins which haven't changed hands since prior to January 13, 2013).  It's important to note that if, say, a wallet address has a 1,000 coins in it then someone spends so much as 0.000001btc out of that wallet, it marks all 1,000 of those coins as 'alive'.  We are looking for wallet addresses which have been completely untouched since prior to January of this year.

You should then be able to do some statistics showing how many zombie wallets come to life over time and make some reasonable predictions about how many are likely to remain forever lost.

Are there lost/zombie coins after January 1 of this year?  Sure, some, but they are probably statistically insignificant compared to all of the lost coins back when they were essentially worthless and people didn't even bother keeping track of wallets.

My argument is this.  Any coins which existed prior to January 13, 2013 are potential zombie coins because they represent a massive multiplier value factor.  Most anyone who owned coins prior that date would have either sold them (to make a 10x profit) or possibly moved them to a newer wallet as a lot of wallet technology and discussion has evolved over this time period.

It would take an enormous amount of willpower not to sell coins which have become worth 10x or more what you originally paid for them or, at least, move them to a more secure wallet technology.

Realize too that I am not simply talking about coins which have been 'cashed out' for fiat; the simple act of moving coins from one wallet to another alone (even if it's a wallet owned by the same person) still shows that these coins are 'alive and well'.

Will we ever guess accurately what percentage of zombie coins are truly dead and which are just in deep storage waiting to come to life again?  I think using some statistical models showing how many come to life over time should allow us to make some informed speculation at least.  My personal guess is that majority of coins that have not changed wallets since prior to January 1, 2013 are likely completely dead/gone/lost destined never to come back to life.  By studying how many of them do rise from the dead, month to month, should give us a reasonable estimate of the rate at which zombie coins rise from the dead.

John

[Birdy]

Yes, a statistical analisys could probably give some insight.

Are there lost/zombie coins after January 1 of this year?  Sure, some, but they are probably statistically insignificant compared to all of the lost coins back when they were essentially worthless and people didn't even bother keeping track of wallets.

A lot of the bigger wallet addresses receive small amounts from other people (often with messages), guess you should try to filter that out.

[Cubic Earth]

Yes, a statistical analisys could probably give some insight.

Are there lost/zombie coins after January 1 of this year?  Sure, some, but they are probably statistically insignificant compared to all of the lost coins back when they were essentially worthless and people didn't even bother keeping track of wallets.

A lot of the bigger wallet addresses receive small amounts from other people (often with messages), guess you should try to filter that out.

Only when coins were *sent* from an address would that prove it was alive.  If significant amounts of coins were still being added to an address, that would suggest it was alive, and perhaps such variables could be included in the model.

John's idea is something I've had in mind as well.  The 10x increase is nothing though, we are already at 140,000x valuation increase from the penny days.

[jratcliff63367]

Yes, the cut-off date is somewhat arbitrary.  Once you have a script that can produce the graph you can try different cut-off dates easily enough.  Also, it's probably better to describe this as a search for dead wallets, not so much dead coins.  How many coins are in all of those wallets that have had zero transactions for a very, very, long time? Inquiring minds want to know.  The 'days destroyed' graph on blockchain really doesn't do much for me.

I just want to know how many dead wallets (and their sum-total coins) are still out there month to month and day to day.  That's a graph blockchain should have.

John

[theonewhowaskazu]

Stealing is stealing, even if it is by 'majority vote.' Hence why the Bitcoin protocol should never be changed to modify the owner/value of coins.

If 60% of the population voted to forcibly confiscate & redistribute the wealth of the other 40% of the population, you think that wouldn't be theft?

If you are using the current version of bitcoin, you can keep using it.  It wouldn't matter if 99% of people decided to use a different version, no one could or would steal your coins, no matter what changes the different version contained.

Now, if you are complaining because 99% of people decide they don't want your coins, and hence, they become almost worthless, well I'm *not* going to let that bother me.

Stealing is one thing, and it is foul; accusing people of stealing (or wanting to steal) just for valuing an asset differently than you might is another, and, also foul.

*edited because I forgot the word *not*

So, let me get this straight.

You're basically attempting to start some sort  of bizarre "movement" to try randomly get everybody to agree to not accept early adopter's coins? WTF? That's supposed to instill certainty? This has got to be the single dumbest thing I've heard on this forum ever. Period.

[justusranvier]

But no one can move the lost coins so their reintroduction into circulation via cryptographic compromise could be uncontrolled and devastating to the Bitcoin economy.

The effect will be underwhelming. The abandoned/lost bitcoins will not be found all at once and the miners who recover them aren't going to immediately turn around and spend the entire find the next day. By that time the economy will be large enough that it only causes a mild effect, if it's even noticeable at all.

Whether it's via some yet-unknown mathematical weakness or quantum computers that makes the private keys recoverable, it will only gradually become possible. So I imagine it will be done by investors combining their resources in order to build machines capable of doing it, much like present pool mining. Any old balances recovered this way will thus not go to a single entity but will get spread out among a large group of shareholders.

You still get too freaked out over the concept of a currency that nobody controls. It's not going to become a problem in practise - modern cryptographic algorithms don't go from no known weaknesses to being trivially broken overnight. There will be plenty of advance notice and bitcoin holders will adapt ahead of time. It will be priced in long before it happens.

[Syke]

I just want to know how many dead wallets (and their sum-total coins) are still out there month to month and day to day.  That's a graph blockchain should have.

It would be dead wrong. You might as well flip a coin to decide if a wallet is lost or not. It would probably be more accurate than using any sort of analysis you could come up with.

[BurtW]

Why another thread?  Why not just add on to one of the dozens and dozens of previous threads that have proposed exactly the same thing.

You can yammer about this all you want.  It is never going to happen.  Mostly because you are trying to fix something that is not broken.  There are many other more pressing real issues that are being worked on.

[Elwar]

I think it could be helpful to remove some of the uncertainty around old coins, to see if they are truly lost or not.  This is an issue that is will never go away, and in fact will only grow more troubling as bitcoin grows in value.  My proposal is to pick a comfortably far off future date (maybe 1 year) and say that any coins on addresses that have been stagnant since some long-ago date (how about Jan. 1st, 2011?) would be removed from the money supply.

If you had coins on an old address, all you would need to do to protect your wealth would be to transfer them to a new address, or just send a single satoshi, before the future cut-off date.  This would be a one-time event that would require a hard fork and therefore a consensus of the community.

Every market thrives on information.  I would argue that large uncertainty about the amount of lost coins is hurting bitcoin adoption, especially its use as a store of value.  I think this proposal would help bitcoin adoption by providing a count of the number of controllable coins, and it would do so in a manner that would not unfairly penalize anyone. 

Perhaps some therapy could help you with your old coin uncertainty.

[gmaxwell]

So, let me get this straight.
You're basically attempting to start some sort  of bizarre "movement" to try randomly get everybody to agree to not accept early adopter's coins? WTF? That's supposed to instill certainty? This has got to be the single dumbest thing I've heard on this forum ever. Period.

Shame on you.  Disagree, by all means— it's a controversial subject.  But at least understand what you're talking about. There is nothing about "early adopter's coins" in his commentary, he's suggesting a requirement that after a long notice period old coins need to move or will become unmovable. It has nothing to do with not accepting early adopter coins, though it would require people with those coins to move them once. If it was done the way he envisioned it working not a single person would lose access to their coins, though some people might be moderately inconvenienced. The effect of it is that ambiguity about how many actual coins exist would be substantially reduced. Maybe this is a terrible evil horrible idea, but at least complain about what he's actually suggesting and not some paranoid delusional parody of it.

But no one can move the lost coins so their reintroduction into circulation via cryptographic compromise could be uncontrolled and devastating to the Bitcoin economy.

The effect will be underwhelming. The abandoned/lost bitcoins will not be found all at once and the miners who recover them aren't going to immediately turn around and spend the entire find the next day. By that time the economy will be large enough that it only causes a mild effect, if it's even noticeable at all.

Whether it's via some yet-unknown mathematical weakness or quantum computers that makes the private keys recoverable, it will only gradually become possible.

Your speculation has higher entropy than my speculation.  Multi-collissions scale much much better than linearly. E.g. even with just the rho method only moderately more computation is required than the 2^128 operations expected to compromise a single ECDSA key to compromises almost all ECDSA keys on a particular curve.

Collision attacks (e.g. all known methods of discrete log solving) favor single large attackers, they are unlike mining is that the probability of success is not constant but goes up the more work they've done.

There will be plenty of advance notice and bitcoin holders will adapt ahead of time. It will be priced in long before it happens.

There will indeed be plenty of notice that a particular scheme is looking weak, and the non-lost coins will be moved... you could have easily been quoting me there on "modern cryptographic algorithms" but I think you're wrong when it comes to practical attacks: The reason we get advanced notice is generally from certificational weaknesses which don't translate into practical attacks. With MD5 collisions we went basically overnight from (I think) two known examples with sensible files existing in the world, to a tool that would produce them freely on my desktop— though we had many years of warning.  Bitcoin further complicates that with single coins being high value: How could an economy functioning on a few thousands BTC "price in" the sudden reintroduction of a single 111,000 BTC "lost coin"?

I don't see how any economy could price that in even if the reintroduction was relatively slow, whatever that would be it would constitute an enormous highly distorting value transfer from the entire economy to whomever operates those cracking devices.  I can't tell you exactly how people would deal with it— blacklisting those coins, or just abandoning Bitcoin— but I can pretty much promise that people will not just tolerate it if such a thing came to pass.

I am not saying that we couldn't absorb the reintroduction of lost coins _today_, I am saying that it's possible that they won't become recoverable until Bitcoin had deflated past any reasonable prospect of being able to accommodate absorbing them.  We might disagree on where that boundary would be, but I don't think you can convince me there there isn't a degree of deflation under which a point recovery of a single gigantic coin is basically a currency extinction event if permitted.

People who understand Bitcoin well enough to be uncertain about the lost coin situation aren't the type of people who are on the fence about adoption. Go ask 1000 people who haven't started using Bitcoin yet, and I bet no one will tell you that their sticking point is the lost coin issue. Either they won't know anything about Bitcoin, or they'll say "it's not backed by anything", "isn't that a pyramid scheme?", etc.

Uh. Dunno who you're talking to, but one of the number one things people say to me is "What happens once all the coins are lost, without inflation you'll eventually run out of coins!"  and then I explain that people can just trade in smaller and smaller increments (adjusting units as they go— millibitcoin etc) and I think about 50% of the time they respond "Hm! interesting!" and about the other 50% of the time they say "whoa whoa! what happens when someone finds grandmas old coins, they'd rule the world!"

(Of course, if you've got someone actually concerned about that, I suppose that it means you've already convinced them that Bitcoin will rule the world, even if its a bad thing... which is perhaps what you meant?)

[theonewhowaskazu]

Shame on you.  Disagree, by all means— it's a controversial subject.  But at least understand what you're talking about. There is nothing about "early adopter's coins" in his commentary, he's suggesting a requirement that after a long notice period old coins need to move or will become unmovable. It has nothing to do with not accepting early adopter coins, though it would require with those coins to move them once.  Maybe this is a terrible evil horrible idea, but at least complain about what he's actually suggesting and not some paranoid delusional parody of it. 

Now, if you are complaining because 99% of people decide they don't want your coins, and hence, they become almost worthless, well I'm *not* going to let that bother me.

He literally just said in his previous post that he doesn't have a problem with trying to make "my" old coins "almost worthless". I have no idea where YOU get you're "paranoid delusional parody" because I'm literally describing EXACTLY what he is proposing.

He wants to convince, 99% of people, that "old" coins are worthless and should not be accepted, because if they were this might cause "uncertainty", presumably in the price. What he's proposing is basically the same style of "money printing" theft Bitcoin is trying to protect against, only in reverse. Rather than trying to devalue other's currency to give yourself more of it, you're trying to destroy other people's currency to increase the value of your own.

[justusranvier]

Bitcoin further complicates that with single coins being high value: How could an economy functioning on a few thousands BTC "price in" the sudden reintroduction of a single 111,000 BTC "lost coin"?

It depends on the velocity of money at the time and how quickly they are spent. If somebody cracks an ancient private key that's equal to a substantial fraction of the previously circulating BTC supply nothing happens right away. It's only when they spend it that anything changes.

Given the risk of somebody suddenly owning decades worth of world economic output all of a sudden, I'd expect that once ancient weak key cracking approached feasibility large numbers of people (millions) would form a pools to crack them that would distribute the recovered bitcoins as dividends.

If it ever turned out to be a huge threat, then the people alive at that time would have a huge incentive to solve the problem.

[tvbcof]

... If it was done the way he envisioned it working not a single person would lose access to their coins, though some people might be moderately inconvenienced. ...

It would be much more than a moderate inconvenience for me.  I always thought it was dumb to have more than a certain relatively small number of coins assigned to a certain address.  Further, I thought it was dumb to have said addresses either easily accessible or accessible in a group. (*)  Tracking down and dealing with my stash in it's totality would be time consuming and difficult, and my methods were designed under a specific understanding of how the system worked.

(*)  "Ever noticed how anyone who drives slower than you is a retard and everyone who drives faster is crazy?"  Computer security is a lot like this

[phillipsjk]

I have a dead wallet because I am more paranoid than average about my computer security.

I still have the private key. Once I set up a secure computer, I plan to move those coins. So far they have been not moving for about 2 months.

1 year's notice is not nearly enough. I *might* agree with a 5 year time-frame suggested by another poster.

Keep in mind that Bitcoin may be declared illegal in your jurisdiction for a period of 10-15 years. Would you really want to risk jail-time (or the death penalty) just to move coins around every year?

[Trongersoll]

Uncertainty is good. Nothing is certain in this world.

[drawingthesun]

Not sure if this has been suggested yet, but how about a system where old coins evaporate over immense periods of time?

So after 100 years 0.1% of the unmoved coins start to evaporate and becomes miners fee.

If you don't want evaporated coins, move them at least once every 100 years. This means over a long enough time line we stick with 21 million coins.

Peoples mistakes (lost coins) become another incentive to mine. Note that only large time scales will be fair, this system takes hundreds of years to erode away even a bit of value.

But taking old coins completely within 100 years? Pure theft, plain and simple.

And 1 year? Whoa that is madness, 1 year is nothing.

[Cubic Earth]

He literally just said in his previous post that he doesn't have a problem with trying to make "my" old coins "almost worthless". I have no idea where YOU get you're "paranoid delusional parody" because I'm literally describing EXACTLY what he is proposing.

He wants to convince, 99% of people, that "old" coins are worthless and should not be accepted, because if they were this might cause "uncertainty", presumably in the price. What he's proposing is basically the same style of "money printing" theft Bitcoin is trying to protect against, only in reverse. Rather than trying to devalue other's currency to give yourself more of it, you're trying to destroy other people's currency to increase the value of your own.

What do you think we are all hear doing?  We have created a currency - Bitcoin - that is devaluing other currencies, and almost every person on these forums hopes that will happen. This is a free market of ideas.  Are we stealing from holders of U.S. Dollars and every other fiat currency?  That would seem to be conclusion your logic leads to.  Nobody is going to want dollars anymore if our grandest bitcoin dreams come true.  Someone is going to left holding the bag.  So whatever you are accusing me of doing to you, you are most certainly doing the same thing to many others.

So again, do you think someone who is trying to design a better currency than ones currently in existence is a thief?  It almost sounds from the tone of your comment that you think you have a right to hold value in bitcoin.  I think you have a right to run whatever kind of software code you, including bitcoin, and a right to hold bitcoins, but the value is conferred by others, by their own free will.  Although perhaps you believe people should be forced into using bitcoin?  I would see that as a greater injustice than our current system.

Furthermore, I knew I should have titled my post 'lost coin uncertainty', and if you had read my idea carefully, you have seen that is exactly what I am describing.  Old is the sense of 'not moved for a long time', not old in the sense of 'when were they mined'.

1 year's notice is not nearly enough. I *might* agree with a 5 year time-frame suggested by another poster.

As for the dates - I just made some up to illustrate my point.  Any timeframe the community could agree on would be fine.  The first date I picked, Jan 2011, had to do with how much the value has increased since then and hence the likely hood that a massive value of coins has been forgotten.  But that first date could be today, or next year.

If it ever turned out to be a huge threat, then the people alive at that time would have a huge incentive to solve the problem.

I agree with this.  But if people have a tendency to imagine threats long before they are real (as we are doing here), and let such knowledge influence their behavior ahead of time.  As time goes on, this problem that we are discussing now will only grow in severity.  It's clear from reading this thread that most of the current participants in the bitcoin community vehemently against this concept.  But I can guarantee you this will change, as new bitcoiners start asking questions about the money supply.  Lost coins aside, I think we all know there is a chance that people will move to a new currency that has less early adopter advantage.  In fact, I perceive that to be the single biggest risk to the value of my bitcoin holdings.  The network effect and the security of the miners is what keeps people in BTC.  Most newcomers certainly don't like the idea that some early folks have tens of thousands of coins that they got "for free".  Of course they weren't free, and they took a risk, saw the future, etc., I am not disputing that, but psychology is just as important as truth in a money system. 

If anything, this proposal seeks to alleviate those concerns of potential newcomers by demonstrating that there is not nearly as much money in the hands of the early adopters as they might fear, and it attempts to do so without depriving anyone of any coins that they control.

Honestly, if I had once had 50,000 coins, and I lost control of 45,000, I would be quite happy to see this idea implemented.  By proving that those 45,000 coins were in fact gone, I would increase the value of my remaining 5,000 coins be more than if people were just left guessing about the 45,000.

[giszmo]

Ok, as there is so much hatred towards the the OP:

I totally agree that unmoved coins should be removed. I disagree with the OP that this should be some singular event (I don't like singular or recurring events like the reward halfing) but it should be some rule like coins that did not move in 10 years are dead.

Yes, I gave bitcoins to friends and family and told them that they just should remember that it's called bitcoin and if in ten years they find the cd and have no clue what bitcoin is they should throw it away. Else, they would be rich. Well, guess what, they are all excited about their bitcoin holdings now and within 10 years they would not miss to move their coins, so I am not worried about people loosing their coins.
I think 10 years would be a reasonable amount of time and being introduced now, it shouldn't take effect before 2023 but I think we should do this because:

We will loose considerable mining capacities to people trying to crack those big addresses that don't move to the next more secure algorithm once the current ones become insecure.
The sins of the past where tons of dust is created will also get wiped out of the block chain once the nodes may forget about +10a transactions.
It is not stealing when people can at no costs keep their wealth as a client of that generation would automate the refresh and the transacitons would be for free as they would not be urgent and certainly qualify for zero fees. Miners not including those keep-alive transactions without a fee would be attacking bitcoin.

The down sides are:
Sending reduces the security by revealing the public key which is why you would better not continue using the address after refreshing it, which would be maybe sad for some vanity addresses that only collected and never spent coins but maybe a rule could be derived that would allow to keep addresses alive without revealing the public key.
Sending requires becoming active somehow and I guess this fact is what really pisses off people here. Hey, I bought a car and now you want to install a dead man switch that I have to press every minute? Yeah, I understand that but when the current algorithms become insecure it will become an issue and I am 100% sure that we will come to some such agreement at some point, so I would rather want it to be in place before it's urgent. Sure, if Satoshi is the GOV he will not want to move his coins never ever before the world agreed to use his coins but hey, in 10 years we most likely have ZeroCoin or something anonymous like that and even if not, it should be possible then to move the addresses without people knowing who did it.