Gmaxwell's “prove how (non)-fractional your Bitcoin reserves are” scheme

https://iwilcox.me.uk/v/nofrac

by sillysaurus2 • 12 years ago

114 76 12 years ago

$Preview of Gmaxwell's “prove how (non)-fractional your Bitcoin reserves are” scheme$ iwilcox.me.uk

The developer is Gregory Maxwell, aka nullc. Here's a very interesting thread in which he proposes that the bitcoin community should demand that every bitcoin exchange (and every other type of service which can hold bitcoin on your behalf, like webwallets) continually prove that they are not fractional reserve. In other words, proof that if every user of the service simultaneously tries to withdraw all of their bitcoin, then the service would be able to honor all withdraw requests: http://www.reddit.com/r/Bitcoin/comments/1yj5b5/unverified_p...

"I think that as a community we should start demanding these services continually prove that they are not fractional reserve. We cannot effectively eliminate the need for trust in these sorts of services, but we can certainly confine the exposure and eliminate a lot of this drama. With Bitcoin it's technically possible to prove an entity controls enough coin to cover its obligations— and even to do so in ways that don't leak other business information, and so we should. But this isn't something specific about MTGox, it's something we should demand from all services holding large amounts of third party Bitcoins. I wouldn't even suggest MTGox should do it first, rather— it sounds like a great move for their competition to differentiate themselves."

Here's the takeaway:

"This would leak the total holdings, and some small amount of data about the number of accounts and distribution of their funds, but far far less than all the account balances. Importantly, though— it could be implemented in a few hundred lines of python."

In case anyone from Coinbase is reading: you have a unique opportunity to be the first webwallet service to implement this, and thereby make the entire bitcoin community instantly fall in love with you. It would also set a minimum standard of quality for webwallet services in general, which would add a lot of value to the bitcoin ecosystem. It seems like this might be a pretty big business opportunity.

mjn• 12 years ago

The developer is Gregory Maxwell, aka nullc.

This guy seems to be everywhere! He's a prolific Wikipedia contributor (administrator + many thousands of edits), and was also the guy behind the dump of a ton of pre-1923 JSTOR documents to the Pirate Bay, which in part helped pressure JSTOR to un-paywall its old/PD articles (http://arstechnica.com/tech-policy/2011/07/swartz-supporter-...).

chris_wot• 12 years ago

The guy was amazing on Wikipedia.

patcon• 12 years ago

Not to discredit the very capable developers discussing this, but in the interest for giving credit where credit is due, didn't Peter Todd suggest this back in his Bitcoin 2013 presentation on off-chain transactions? I seem to remember him explaining something similar on a rooftop patio in Toronto last spring after a Bitcoin Toronto meetup.

EDIT: http://www.youtube.com/watch?v=4d3LA8KpdMQ#t=6m45s

nullc10• 12 years ago

I believe this was most extensively discussed as part of a long chat that Peter Todd was a part of, so no surprise that you've seen him talk about it. Off-chain banks stuff has been a long term pet interest of his.

In that discussion we applied a merkel-sum tree data-structure— a pet datastructure that I'd previously proposed for making compact proofs of blockchain invalidity in Bitcoin (in order to make a future bitcoin world where no one runs full nodes safe from inflation and theft by miners)— to PT's bank fraud proofing application.

You may find the log interesting: https://people.xiph.org/~greg/bitcoin-wizards-fraud-proof.lo...

Search for "auditable off-chain transactions" and "Merkle-sum-tree"

(I left in a lot of unrelated stuff since it makes the meandering conversation make a bit more sense. Though a lot of this continues a long running dialog about cryptographic-wankery that has been going on for years)

Ultimately these schemes require the use of a jamming free broadcast network of some kind... otherwise they run into the same problems certificate transparency has where you can substitute the commitment on the fly. Fortunately, Bitcoin provides a global consensus mechanism which could be used to directly attach the commitment to the coins being spoken for.

petertodd• 12 years ago

I'm pretty sure I remember either Gregory Maxwell or Andrew Miller suggesting it to me first, and I think it might be the latter's idea originally. (at least in the Bitcoin space) Andrew has done a lot of work on "merkelizing" data structures: http://www.cs.umd.edu/~amiller/gpads/

higherpurpose• 12 years ago

Fractional reserve? I don't like that. It's like building a house of cards or a ponzi scheme. You shouldn't be able to say you have 10x of the value you actually have.

jeremyjh2• 12 years ago

And that is the entire purpose of the proposal.

minimax3• 12 years ago

Right? Banks making loans? It's preposterous.

waterlesscloud• 12 years ago

Fractional reserve is sort of a misappropriated term as applied to bitcoin exchanges. Such exchanges aren't making loans, so where would the money be going? There's no good reason for them not to have 100% of the funds in reserve.

dmm• 12 years ago

Full reserve banks are possible. You just have to maturity match everything, meaning every loan of term t has a matching deposit with a maturity of t. Deposits in demand accounts could not be lended.

I have no idea whether this is a good idea or if it would work as a business.

splintercell• 12 years ago

Full Reserve banks don't make loans?

Fractional Reserve banks enable you to have your money and be loaned out at the same time(thereby stretching the money supply like an elastic rubber band).

Full Reserve banks ONLY loan out money which was specifically deposited to them for the purposes of being able to be loaned out.

ef42• 12 years ago

Sigh. It's a fallacy that making loans implies fractional reserves as normally understood.

Banks can still make loans simply by offering certificates of deposit. This is the above-board way of loaning out people's money -- you make it absolutely clear that taking it out early has a cost, because the money is locked up in (hopefully) profitable ventures.

Would that be less profitable for the banks? Not really -- they would just adjust their prices to compensate, by charging fees on idle money that's instantly redeemable.

And if you let a secondary market for CDs flower, customers can still get good liquidity. Just in a way that's better subject to market discipline.

hayksaakian• 12 years ago

Not sure if joking

You do realize that modern finance depends on this notion?

For example, Banks need no more than 10% of a loan as cash on hand.

baddox• 12 years ago

You do realize that modern finance depends on this notion?

Modern finance also causes some very serious societal problems, in my opinion.

Coffeewine• 12 years ago

Alternatively, a fractional reserve institution can offer interest on accounts which might entice some people to accept the risk of a bank run. Given the popularity of some explicit ponzi schemes in the past month, It seems fairly clear there are plenty of folks who would accept those terms.

infruset3• 12 years ago

At first I was worried of what would happen if the exchange introduced fake nodes with negative balances at the bottom of the tree, but there would be no way for them to hide that without the first real customer up to the root finding out (there would have to be a negative node that he/she could see). This sounds like a great idea!

nullc10• 12 years ago

Unless the negative valued customer and the surrounding customers never logged in... But thats a limitation of the scheme that can't be avoided. If a user never logs in you could just steal just their balance (and correctly set it to zero).

You also must make sure that all customers are seeing the same root, and that you can't do funny business like constantly update it to swap out which customers you're robbing. (e.g. it should be a daily or weekly updated thing).

infruset3• 12 years ago

You've got a point, the root of the tree could be made available to the main charting sites.. or even weekly written into the blockchain.

As for the negative values, I wasn't thinking of robbing anyone, but just pretending you are solvent when really you're not. I'm not sure I see what you mean by "swap out which customers you're robbing", could you expand?

minimax3• 12 years ago

This doesn't give you a way to validate your dollar deposits. In other words a dishonest exchange operator could misappropriate your dollar deposits and this scheme wouldn't tell you anything about it.

teraflop• 12 years ago

The hash-tree scheme described here would work equally well for non-Bitcoin currencies, if I'm understanding it correctly. The only thing missing is the ability to prove ownership of the actual funds backing that tree. So what you're really complaining about is that there are no banks that offer digitally-signed attestation of account balances.

jnbiche5• 12 years ago

no banks that offer digitally-signed attestation of account balances.

Why is this? Seriously, that alone could prevent so much fraud and misuse of funds. Every public company could have a digitally-signed bank balance, updated in real time.

nullc10• 12 years ago

Yup. This is an example of how Bitcoin is superior to USD— USD is not so readily subject to cryptographic proof. :)

You could use it to show that USD obligations jive with third party audits, insurance, or accounts in a bank if you could get the bank to produce signed attestations... though the trust isn't eliminated there, just shuffled around.

pedrohrcunha• 12 years ago

sweet!

already pleading it to brazilian exchanges.

kumarski• 12 years ago

What bitcoin exchanges do HN readers trust?

I've been using http://coinmkt.com

I regrettably used MtGox.com. I'm kicking myself now.

pmorici5• 12 years ago

If you want to buy coins to use I'd go with Coinbase at least to get started. If you want to day trade Kraken looks really promising if they are supported in your jurisdiction. Bitstamp has a decent track record though their local give me pause. Coinsetter is also pretty well put together if all you want to do is pair trade but you can't do true exchange on there.

I've tried coinmkt but I don't like it. Their fees aren't great their deposit and withdrawal methods are limited and there are fees on deposits and withdrawals at least there were when I gave it a try.

jnbiche5• 12 years ago

Bitstamp has a decent track record though their local[e] give me pause.

I'd say Bitstamp has a very good track record, and what's wrong with Slovenia? It's probably about on the level of the Czech Republic in terms of economic freedom, development, level of corruption (relatively low), output, business practices, etc. Would doing business with a Czech company make you nervous?

In business culture, Slovenia looks toward Germany more than toward the former Eastern Block (of which it was never a part).

I mean, Bitstamp's owners are public people. I feel pretty confident that they're not going to run off with their depositors money. Is there something else that you're concerned about?

But I agree, if you want to just buy coins, Coinbase is a good start. And Kraken is looking very good, too, particularly if you're a serious trader (that's who they appear to target).

slg• 12 years ago

I would be a little cautious with Coinbase. There was a recent post on HN[1] in which someone had a 5 figure transaction be approved on the site, but never received his funds. The slow response from Coinbase wasn't too encouraging, but the more troubling issue is that a large sum of money like that can just appear in the system without anyone noticing. It indicates that their level of accounting and auditing isn't up to the standards of even most brick and mortar businesses let alone a financial institution.

[1] https://news.ycombinator.com/item?id=6929705

oznathan• 12 years ago

Bitstamp have proved their good intentions and competence. They also have the most to lose at this point as the market leader. Had zero problems with transfers from there, very quick support replies, seem to be in good relations with their bank. I trust them. Karaken also look promising, but it's a new exchange.

Anyway, best bet is to use several exchanges to take advantage of arbitrage opportunities and to distribute the risk.

roymurdock• 12 years ago

Coinbase. They've got $31m in combined funding from Y-Combinator and Andreesen Horowitz, among others. I like to place my bets where veteran investors have some cash and reputation on the line.

tlrobinson3• 12 years ago

Part of the problem is leaving funds laying around in exchanges, when it's trivial to transfer to your own wallet (though admittedly non-trivial to secure it, it's a tradeoff).

If you quickly transfer in, exchange, and transfer out you don't need quite as much trust.

Himself• 12 years ago

I've used coinmkt more and more in the past week. I like them, seem to still be working things out, but if you trade multiple currencies they're a good bet. Being based in the USA doesn't hurt either.

jrockway• 12 years ago

It turns out that MtGox used all the deposits to buy Magic cards. They now have the world's most excellent cube. We all should have seen this coming.

M4v3R2• 12 years ago

Or, just use a system like we use on Bitalo, where fractional reserves are impossible because of use of multi-signature Bitcoin addresses, which means funds are specifically tied to user wallets and exchange operators cannot use them without user's signing all transactions by himself.

stingraycharles• 12 years ago

Exactly, this is specifically where that feature is designed for: it makes you your own arbiter of what an exchange wants to do with your money.

wmf• 12 years ago

It's not clear that day traders are willing to pay the fees to put their transactions on the blockchain.

M4v3R2• 12 years ago

They are paying much higher fees for trading on "normal" exchanges. Transaction fee is only 0.0001 BTC, normally you pay WAY more for trading on Mt. Gox/Bitstamp.

nullc10• 12 years ago

This requires one transaction per trade, effectively.

While systems like that have many applications— and should be used where they can, they aren't a replacement for large scale markets like MTGox or for ultra-low-cost instant payment systems.

gnaritas3• 12 years ago

The blockchain can't handle the transaction volume of currency exchanges; that just won't work.

mynewwork• 12 years ago

Wait, the blockchain can't handle the transaction volume of the trades happening today?

What happens as businesses (overstock, etc) start accepting bitcoin? Will bitcoin never be able to handle to volume of an amazon or walmart?

snake_plissken• 12 years ago

Can't all of this information be found in the block chain if you know the addresses the exchanges are using?

nullc10• 12 years ago

No. In theory if you know all the addresses you know how many coins they have— but to know the exchange is not fractional you must also know something about its obligations.

aston• 12 years ago

Say you were a shady Bitcoin banker with 5000 BTC in deposits, and you wanted to steal 1000 while still looking like you're on the up-and-up by implementing this idea.

First, you announce that you only have 4000 BTC in deposits. Then you build this tree, and at the very bottom layer you add a node with a -1000 balance. You pair that node with your (or a conspirator's) real node holding more than 1000 so that any node above yours (read: everyone else) sees a positive balance at every point in the tree. Everyone can verify they're in the tree, the numbers add up to what you claimed publicly, but you're now successfully running a fractional reserve! And the only way to uncover such a scheme would be to publish all of the balances for every account.

Am I missing something?

Edit for clarity: the node you pair with is your own, so that no real user sees the negative sum.

tlrobinson3• 12 years ago

Discussed here: https://news.ycombinator.com/item?id=7278325

vbuterin4• 12 years ago

Suppose the balance sheet is:

    [ -1000, 1000, 2000, 2000 ]

The Merkle tree is:

    [ -1000, 1000, 2000, 2000 ]
    [ 0, 4000 ]
    [ 4000 ]

You actually owe 5000 BTC, but it seems like you owe 4000 BTC. Seems so far so good. The problem is, what happens if you try to take advantage of this opportunity.

Case 1: other people withdraw first.

    [ -1000, 1000, 0, 0 ]
    [ 0, 0 ]
    [ 0 ]

Nobody knows that anything nefarious has gone on. However, everyone else has successfully gotten their money out so you've actually defrauded no one.

Case 2: you withdraw first.

    [ -1000, 0, 2000, 2000 ]
    [ -1000, 4000 ]
    [ 3000 ]

Now, the other 2 users actually can see that something is wrong, because the Merkle branch will have a -1000 BTC node sticking out.

So in theory, as long as there exist users who don't check their Merkle branches, and those users are identifiable, it probably is possible to run a slight fractional reserve undetected. So the protocol is suboptimal. But it's not really "broken". I do wonder if it can be improved though, perhaps with some kind of ZKP protocol.

nullc10• 12 years ago

Oh sure, you can sum and compare the balances under ZKP and even hide the total amount. But the problem is that as soon as you invoke a ZKP for general computation you take into the realm of barely practical moon math.

... And you still don't fix the problem that balances which are unchecked can be diverted.

In the IRC log I posted I went on to suggest that a service could have a rule that _permitted_ them to take your balance if you don't check it periodically— e.g. they could just withdraw it into their own pocket. You could prove you checked it (or that you tried and they wouldn't let you). By doing so you'd actually create a real incentive for people to check, though I suspect boobytrapped balances wouldn't be very welcome.

Regardless— it still confines the extent of fraud that is possible.

tlrobinson3• 12 years ago

The big problem with this is convincing businesses to publicize their total customer deposits, which is extremely interesting information to competitors.

Though it could be a good way for new/small exchanges to differentiate themselves and gain trust of the community, which could force larger and larger exchanges to do the same until it's common practice (as mentioned has happened with provably-fair gambling sites)

Sambdala2• 12 years ago

If anyone is interesting in helping, I'm going to spend my evening trying to implement this here: https://github.com/ConceptPending/proveit

My email is in my profile, and I'm happy to Skype chat with anyone who wants to help.

Sambdala2• 12 years ago

The basic implementation is now complete.

I'll flesh it out a bit better tomorrow.

Gmaxwell's “prove how (non)-fractional your Bitcoin reserves are” scheme

Related Stories