Interest in a P2P Exchange

[joesmoe2012]

I don't see how an Alt-coins exchange would help. The problem isn't getting BTC in particular, its getting ALT-coins in general.

[toffoo]

Well, I suggest the idea because it would be so much easier and more feasible to do than a distributed BTC vs. fiat exchange.

There is already a well established market for trading BTC vs. Altcoins (centralized of course, at BTC-e and a few others) so this would be an excellent proof-of-concept that a distributed P2P exchange can work and thrive.

I can envision some clever developer coming up with an open-source distributed exchange client that could perhaps cryptographically escrow your BTC and LTC holdings, maintain a distributed order book, and then instantly swap the BTC for LTC when a trade has been agreed, without counterparty or clearing risk.

Something like that would probably catch on.

[bitbitcoins]

look at i2p technology perhaps ?

[Bitcoiner_cph]

I have been giving this topic some serious thoughts. I have some general ideas about how to solve the problem with making a p2p exchange.

My ideas are not that concrete yet, but maybe they den be more clear if I share my thought here and we develop ideas here.

It should be fairly easy to make the bitcoin par of the p2p exchange, the hard part is moving the fiat.

1.
One way to move the fiat is by letting people transfer to each others bank accounts, "anonymously" or covered by another "word in the messages to the bank, maybe an code". Okay this should also be easy to make. But it would be very easy for people to abuse the system, take the BTC and never do the wire, right? So the p2p network have to include a WOT function, then the network could be build slowly with smaller transactions between people that have some real life trust, and when the WOT develop, more and more people could enter - and the p2p would expand.

I think this kind of p2p exchange would require a system where new people are introduced by trusted people in the network, and only with smaller amounts until they build their trust WOT. The WOT had to be managed though a new protocol that is included in the p2pexhcnage, I cannot see how this should be done in practice at the moment. Will this idea work?

2.
Making a peer to peer protocol that serve al transactions, but make smaller centralized privately owned clearing houses, that make the actual fiat transactions in one or several countries. These would be small, un maybe undercover/semi-anonymous (?), and there would have to be a WOT rating system for these smaller clearing houses?.. they have to operate in a way so they can be opened and closed fast, in order to keep regulators away.

3.
Drop all official exchanges of bitcoins and just make a pricing system, that people that trade in person via OTC and bitcoin-otc could use to report trades??

4.
Make an exchange that uses NASH EQUILIBRIUM AND MUTUAL ASSURED DESTRUCTION. It can be made via TOR and therefore anonymous, it hold the future of bitcoin exchange IN MY BELEIVE!!! ( I am a programmer and I can make this exchange 100%... I need designer and a TOR hosting expert!!! (write me a PM).

http://www.libertariannews.org/2013/05/03/using-the-nash-equilibrium-and-mutually-assured-destruction-to-by-pass-currency-exchanges/

[BlueNote]

4.
Make an exchange that uses NASH EQUILIBRIUM AND MUTUAL ASSURED DESTRUCTION. It can be made via TOR and therefore anonymous, it hold the future of bitcoin exchange IN MY BELEIVE!!! ( I am a programmer and I can make this exchange 100%... I need designer and a TOR hosting expert!!! (write me a PM).

http://www.libertariannews.org/2013/05/03/using-the-nash-equilibrium-and-mutually-assured-destruction-to-by-pass-currency-exchanges/

That sounds crazy and drastic to me. Both people "destroying" a good amount of value in BTC over a simple trade? How is that an improvement on escrow? An arbitration/escrow service or WoT (web of trust) can facilitate trade without having to cause both people to lose and send their bitcoins into a black hole.

There are real issues that may come up in a trade where both people may have valid grievances against the other. This "solution" has them dumping twice the value of the trade down the tubes. And then what - they walk away seething with anger and never trade again? "Mutually assured destruction" has to do with psychopathic governments exchanging nukes with each other. Trading is supposed to be the opposite of that; it's supposed to be civilized.

Why would anyone risk 20 BTC to trade 10 when they could just pay a small fee to an escrow service?

[charleshoskinson]

Because you have to trust somebody

[bytemaster]

Because with an escrow service you have to trust a 3rd party AND if there is a conflict then that 3rd party has to judge the impossible:   'was it damaged before, durring, or after shipping?'.   What were the original promises?  Etc.

With NashX both parties have a powerful incentive to work things out.  

Take child support disputes... it is often better to settle and pay 'more' or 'less' than you think is fair than to fight it in court.  The high risk, the amount you lose in legal fees is greater than the amount you hope to gain, etc.   NashX eliminates the need for a 3rd party arbitrator.  

There is only one downside to NashX's approach... it can be a game of chicken and someone who is 'crazy' can win concessions in their favor.   My ex-wife does this to me all of the time.     We both 'lose' by going to court and we both 'gain' by staying out of court.   If we simply 'split' things then we would each see $2000 savings by avoiding court.   But she decides she would rather loose that $2000 than see me gain $2000 and therefore demands the 'savings' be split $3000/$1000 in her favor or else we go to court and we are both out $2000.

So if you are 'crazy' you enter into a deal with someone and lock up 40 BTC... then they send you 10 BTC payment and you 'default' and don't send them the goods worth 10 BTC.    You then tell them that they can settle the dispute and get their 20 BTC back for a net loss of 10 BTC or they can lose 30 BTC.    Given the choice between losing 10 and losing 30 it is only 'rational' to lose 30 *if* you expect to gain something for that 20 BTC.   If the money goes into a 'black hole', the only thing you gain is 'punishment' of the bad guy.    This only works if most people agree to punish the bad guy AND the bad guy wants to repeat the scam, but this is a prisoners delimia because everyone has incentive to 'defect' and take the 10 BTC loss instead of 30 BTC loss even though everyone is better off cooperating because it would eliminate the fraud in the first place.

As a 'crazy bad guy' you decide to play this scam on 10 people.   If 30% of the time people destroy the funds, and the other 70% they agree to settle... then you net 70 BTC from those you scammed and lost 60 BTC from those who opt for mutually ensured destruction causing your whole scam to net 10 BTC.  

Therefore the amount that individuals are required to risk depends upon the average 'defect' rate in the prisoners delimia of accepting a 10 or 30btc loss for the benefit of others.
 


    

 

[BlueNote]

I would rather build a network of people who are accountable to each other. I think the promise of P2P exchange (not just buying and selling bitcoin) is that you can build a functioning society that parallels the real world. I don't see why you can't have trusted individuals/institutions and a working legal system in the digital realm if you have the right design.

The quest for total anonymity doesn't strike me as realistic. It makes sense for some things, but I don't think it makes sense overall in a system of exchange. A strong network of people engaged in free trade is more interesting to me than total anonymity.

Trust is not a big deal if you have the right rules and mechanisms in place to govern people's behavior. Going for total anonymity takes you a different direction where you lose so much of what makes society functional and interesting. I can see some anonymous elements in a P2P network, but I can't see it as a paradigm for the whole system of exchange.

[charleshoskinson]

Do we want paypal or do we want cash?

[waxwing]

Order book functionality and front end websites are trivially implemented.

The problem is and always be: how do you handle the movement of non-bitcoin currencies?

That is not a problem that can be solved with software.

That's what I'm getting at. It seems to be the missing link in the entire puzzle

I agree. The below is a post I made on reddit two weeks ago. It didn't get much traction but I still think it might be an important idea. (EDIT: not sure if my idea here is any big improvement over ideas already mentioned, but maybe still worth a bit of attention, so I'll leave it here).

http://www.reddit.com/r/Bitcoin/comments/1doeih/idea_for_solving_the_exchange_problem_just_add/

(text for lazy):
Consider:
A set S of people possessing USD bank accounts AND BTC balances. They also should have internet banking to make things a lot easier. They all run peer to peer software P.
A new user X wants to exchange USD for BTC. She downloads and runs P.
P selects a member of S (according to some simple criteria e.g. their current BTC balance, and that they haven't transacted recently, etc.).
Member S1 is chosen and instructed to put $Y equivalent into escrow in BTC. Exchange rate set by some reference rate has to be agreed by both parties before continuing.
When X sees that $Y equivalent is in escrow, she wires $Y in USD to S1's account as instructed by software. Note that this is an individual-individual wire transfer and so is less likely to get flagged than a transfer to a known BTC exchange.
The beauty of this approach is that we can obfuscate further by having the software create multiple transfer paths. E.g. Buyer X sends wire to S2 who then wires to S1. 2 of 3 escrow for both steps. Process takes longer (not necessarily a long time in one country, though, in some places it's very quick) but after all we're talking about just initial funding of BTC. Also, creating multiple steps is a hassle and is probably not needed.
A small economic incentive for being an intermediary (members of S) is highly desirable, for obvious reasons. But don't make it too high since the process only has value by being distributed.
Advantages of this approach should be obvious - no centralising of the absolutely critical point of failure in the current system: fiat into BTC. No need for in person contact and no paper/info trail (use encryption) except occasional wire transfers which have no pattern that the regulators could identify. No need for a centralised trading house exposed to DDOS and just general system failure. And even if you want to put 20K USD into BTC you can do it 1K at a time, always wiring to a different person.
It works because it combines the power of decentralisation with the type of transaction that the government cannot possibly regulate away: individual people will always be allowed to give each other small amounts of money or the entire society would collapse!
Finally about escrow: m of n type escrow could be implemented again using the peer to peer software - e.g. a number of escrow third parties (to hold the locked addresses) could be set up and chosen at random, again they would be given a small economic incentive do so, note that they would never have even the ability to take your money in this architecture, they could only destroy it, for which they have zero incentive. Actually I am not very knowledgeable about these bitcoin escrow processes but I know they have been used already.

[QuantumQrack]

Order book functionality and front end websites are trivially implemented.

The problem is and always be: how do you handle the movement of non-bitcoin currencies?

That is not a problem that can be solved with software.

That's what I'm getting at. It seems to be the missing link in the entire puzzle

I agree. The below is a post I made on reddit two weeks ago. It didn't get much traction but I still think it might be an important idea. (EDIT: not sure if my idea here is any big improvement over ideas already mentioned, but maybe still worth a bit of attention, so I'll leave it here).

http://www.reddit.com/r/Bitcoin/comments/1doeih/idea_for_solving_the_exchange_problem_just_add/

(text for lazy):
Consider:
A set S of people possessing USD bank accounts AND BTC balances. They also should have internet banking to make things a lot easier. They all run peer to peer software P.
A new user X wants to exchange USD for BTC. She downloads and runs P.
P selects a member of S (according to some simple criteria e.g. their current BTC balance, and that they haven't transacted recently, etc.).
Member S1 is chosen and instructed to put $Y equivalent into escrow in BTC. Exchange rate set by some reference rate has to be agreed by both parties before continuing.
When X sees that $Y equivalent is in escrow, she wires $Y in USD to S1's account as instructed by software. Note that this is an individual-individual wire transfer and so is less likely to get flagged than a transfer to a known BTC exchange.
The beauty of this approach is that we can obfuscate further by having the software create multiple transfer paths. E.g. Buyer X sends wire to S2 who then wires to S1. 2 of 3 escrow for both steps. Process takes longer (not necessarily a long time in one country, though, in some places it's very quick) but after all we're talking about just initial funding of BTC. Also, creating multiple steps is a hassle and is probably not needed.
A small economic incentive for being an intermediary (members of S) is highly desirable, for obvious reasons. But don't make it too high since the process only has value by being distributed.
Advantages of this approach should be obvious - no centralising of the absolutely critical point of failure in the current system: fiat into BTC. No need for in person contact and no paper/info trail (use encryption) except occasional wire transfers which have no pattern that the regulators could identify. No need for a centralised trading house exposed to DDOS and just general system failure. And even if you want to put 20K USD into BTC you can do it 1K at a time, always wiring to a different person.
It works because it combines the power of decentralisation with the type of transaction that the government cannot possibly regulate away: individual people will always be allowed to give each other small amounts of money or the entire society would collapse!
Finally about escrow: m of n type escrow could be implemented again using the peer to peer software - e.g. a number of escrow third parties (to hold the locked addresses) could be set up and chosen at random, again they would be given a small economic incentive do so, note that they would never have even the ability to take your money in this architecture, they could only destroy it, for which they have zero incentive. Actually I am not very knowledgeable about these bitcoin escrow processes but I know they have been used already.

I like it.. a lot.  Can somebody code this please? 

[charleshoskinson]

Now this thread is cooking. Keep it coming ya'll

[solex]

Now this thread is cooking. Keep it coming ya'll

Agreed. Waxwing is describing an ideal high-level solution.

[BTCLuke]

I agree. The below is a post I made on reddit two weeks ago. It didn't get much traction but I still think it might be an important idea. (EDIT: not sure if my idea here is any big improvement over ideas already mentioned, but maybe still worth a bit of attention, so I'll leave it here).

http://www.reddit.com/r/Bitcoin/comments/1doeih/idea_for_solving_the_exchange_problem_just_add/

(text for lazy)...

This is very similar to what I had in mind as well. I have thought about it quite a lot over the last month and I'm pretty close to cracking this thing, but like everyone else here says, handling fiat is the killer.

While your version of the solution gives a basic premise on what to do with the fiat, I don't see a way to replace fiat transfer functionality for anything faster than an OTC service... Clearly this escrowing of the fiat (specifically the funding of the escrow from a bank or paypal account) is not rapid enough for this system to work in real time with 100s of trades a minute like MtGox does.

And certainly no pretty real-time graphs of the trades like the MtGox API allows for, if I'm not mistaken.

I haven't solved this part yet either, but I've got some very interesting ideas on how the escrow and decentralized order book would work exactly. It's even self-promoting in that just keeping the client open and online will make the owner money automatically without lifting a finger!

Sadly, the best idea I've heard so far for a realistic solution to the fiat problem is Bitcoiner_cph' Nash Equilibrium idea above, although that was ironically suggested as an idea for use with bitcoin, not fiat.

Bitcoin doesn't ever need such a device... Escrows can be funded very quickly since the value itself moves, and it's all just one movement. Fiat, on the other hand, moves an "IOU" first followed by actual funds through a government-approved system. We have no problem coding an IOU mover, but moving the value outside of the guvmint's wires is quite a challenge.

What we really need is someone who has worked with ACH, Wiretransfers, Swift & SEPA systems, and any other banking protocols that truly understands how VALUE is moved between accounts, not just "IOUs."

Once we can duplicate the movement of a fiat's value between two peers without a centralized source like a bank in the middle, then we'll be able to hold that value (and therefore the fiat money itself) in online wallets that we trade in real-time with cryptocurrencies.

[BTC Books]

I agree. The below is a post I made on reddit two weeks ago. It didn't get much traction but I still think it might be an important idea. (EDIT: not sure if my idea here is any big improvement over ideas already mentioned, but maybe still worth a bit of attention, so I'll leave it here).

http://www.reddit.com/r/Bitcoin/comments/1doeih/idea_for_solving_the_exchange_problem_just_add/

(text for lazy):
Consider:
A set S of people possessing USD bank accounts AND BTC balances. They also should have internet banking to make things a lot easier. They all run peer to peer software P.
A new user X wants to exchange USD for BTC. She downloads and runs P.
P selects a member of S (according to some simple criteria e.g. their current BTC balance, and that they haven't transacted recently, etc.).
Member S1 is chosen and instructed to put $Y equivalent into escrow in BTC. Exchange rate set by some reference rate has to be agreed by both parties before continuing.
When X sees that $Y equivalent is in escrow, she wires $Y in USD to S1's account as instructed by software. Note that this is an individual-individual wire transfer and so is less likely to get flagged than a transfer to a known BTC exchange.
The beauty of this approach is that we can obfuscate further by having the software create multiple transfer paths. E.g. Buyer X sends wire to S2 who then wires to S1. 2 of 3 escrow for both steps. Process takes longer (not necessarily a long time in one country, though, in some places it's very quick) but after all we're talking about just initial funding of BTC. Also, creating multiple steps is a hassle and is probably not needed.
A small economic incentive for being an intermediary (members of S) is highly desirable, for obvious reasons. But don't make it too high since the process only has value by being distributed.
Advantages of this approach should be obvious - no centralising of the absolutely critical point of failure in the current system: fiat into BTC. No need for in person contact and no paper/info trail (use encryption) except occasional wire transfers which have no pattern that the regulators could identify. No need for a centralised trading house exposed to DDOS and just general system failure. And even if you want to put 20K USD into BTC you can do it 1K at a time, always wiring to a different person.
It works because it combines the power of decentralisation with the type of transaction that the government cannot possibly regulate away: individual people will always be allowed to give each other small amounts of money or the entire society would collapse!
Finally about escrow: m of n type escrow could be implemented again using the peer to peer software - e.g. a number of escrow third parties (to hold the locked addresses) could be set up and chosen at random, again they would be given a small economic incentive do so, note that they would never have even the ability to take your money in this architecture, they could only destroy it, for which they have zero incentive. Actually I am not very knowledgeable about these bitcoin escrow processes but I know they have been used already.

This is really quite good.  Here's a missing piece:

Integrate knowledge of ACH and SEPA into the routing scheme of the (P) client.  Use bitcoin transfers to hop between the two when necessary, so that wire transfers are never needed.  Then you've got the whole world into the system with no transaction having a base cost of more than a dollar.  What individual members of set S choose to charge is up to them, of course.

EDIT:  I would participate in this, and contribute to a kickstarter-type campaign.

[rebuilder]

Is it yet possible to make split-key wallets? I mean a wallet with a private key derived from keys provided by two or more parties, with each party only knowing their own portion of the key, and the generated private key accessible only to someone knowing all of the input keys? If this can be done, no trusted 3rd party is needed. No possibility for arbitration either, though.

[waxwing]

Nice to see some good discussion going on this.

While your version of the solution gives a basic premise on what to do with the fiat, I don't see a way to replace fiat transfer functionality for anything faster than an OTC service... Clearly this escrowing of the fiat (specifically the funding of the escrow from a bank or paypal account) is not rapid enough for this system to work in real time with 100s of trades a minute like MtGox does.

I had in mind escrow for the BTC. That is to say, the seller of BTC places them in 2 of 3 escrow, waiting for the wire transfer to arrive.

And certainly no pretty real-time graphs of the trades like the MtGox API allows for, if I'm not mistaken.

I hadn't thought about that part at all, but there's no reason why the software can't collect the price data from all the transactions. Clients would probably want to use a reference rate from some external exchange, but they could just as easily use the last price transaction within the P2P network itself, or just agree on their own rate.

A couple more ideas: what I was proposing is only different from existing OTC approaches (including localbitcoins) in 2 ways: first, if it works as designed people can do the transactions without leaving their home (I assume others can wire money from internet banking to a not-pre-defined other bank customer, as I can, using my bank's 2 factor authorization process - but doubtless this must vary a lot from bank to bank) and second, the P2P software can have a lot of intelligence in its algorithm to choose a counterparty/counterparties. If it can be made to work, therefore, this approach is much better than localbitcoins.

It doesn't have to be wire transfers, but it needs to be something with high irreversibility, realistically. Someone on reddit suggested that SEPA transactions are reversible in a certain time frame (I had thought that they weren't). That would limit the speed of these transactions to that timeframe. Does anyone else have info on reversible wires? As for ACH, I only know it exists, not being American I have no knowledge of the nuances.

I'd also emphasize that we should be clear what we're trying to solve. From my perspective, this is not an attempt to create the long term P2P trading structure that might be needed for cryptocurrency based economy in general (as for example Ripple type things are doing now, although I'm in the anti-Ripple camp). This is just about creating a structure to "hack" the banking system such that it's impractical to stop the flow of money into Bitcoins initially.

[edmundedgar]

I don't see how an Alt-coins exchange would help. The problem isn't getting BTC in particular, its getting ALT-coins in general.

It's one of two things that would help mitigate the problems. The other is an alt-coin that controls its own money supply to keep its value pegged to a fiat currency - a kind of parallel crypto-dollar.

That wouldn't solve the problem, but it would help because:
- You'd have functioning price discovery even when the non-p2p exchanges were borked.
- You'd be able to speculate and hedge against exchange rate risk between Bitcoins and the equivalent of dollars, in a censorship-resistant, frictionless environment.
- People could sell stable crypto-dollars for cash in a lot of environments where they can't sell volatile Bitcoins. For example, a lot of shops near me sell online gaming tokens, but it's hard to see them selling Bitcoins because they'd have to keep changing the prices. These could then be traded for Bitcoins on a p2p exchange.

[waxwing]

Is it yet possible to make split-key wallets? I mean a wallet with a private key derived from keys provided by two or more parties, with each party only knowing their own portion of the key, and the generated private key accessible only to someone knowing all of the input keys? If this can be done, no trusted 3rd party is needed. No possibility for arbitration either, though.

I believe that's functionally equivalent to 2 of 2 escrow. Have a look at https://bitcointalk.org/index.php?topic=49370.0

EDIT: I see no reason why we couldn't give the user the choice of 2 of 2 or 2 of 3, with the former being slightly cheaper, but having the risk of your counterparty dying or being irrational enough to destroy their reputation (and receiving nothing) by not paying you. A reputation system could make that risk very small indeed; I could well imagine choosing that option in many scenarios, and yes it would be more convenient.

[solex]

It's one of two things that would help mitigate the problems. The other is an alt-coin that controls its own money supply to keep its value pegged to a fiat currency - a kind of parallel crypto-dollar.

That could be Liberty Reserve? Perhaps that is the best quasi-crypto version of the dollar (also a euro version exists, I think).