IDEA: Unified Bitcoin Exchange

[runeks]

I just had an idea that I think could help spread out trading over more exchanges, offering better prices for selling and buying Bitcoins, and making trading itself more resilient to single exchanges going down.

The idea is that a service, let's call it Unified Bitcoin Exchange (UBE), has accounts on all the major exchanges. It would have to trust these exchanges, because it would need to have enough currency - both Bitcoin and traditional currency - in each exchange account to allow multiple users to - in essence - trade using this account. Not directly of course. The idea would be that instead of going to Mt. Gox, you create an account at UBE and log in here. You then deposit money to UBE like you would at any other exchange and start trading. The "current ask price" in UBE would be the exchange offering Bitcoins for the least amount of money. It could be Mt. Gox, Tradehill, Intersango, or some other exchange. UBE automatically finds the best price on all enabled exchanges. The "current bid price" would be from the exchange that offers the most money to buy Bitcoins. When you place an order, UBE executes it on the exchange that has the best offer, possibly executing half of the order on one exchange and half on another to get the best price. No reason to go to Mt. Gox or Tradehill anymore to trade, just file your order with UBE and it finds the best price available on any exchange.

I haven't quite figured out, when placing bids and asks, if it would be possible to place the bid or ask on all the available exchanges, and when the total of the bid or ask is met - either at a single exchange or small parts of the order spread out over multiple exchanges - it is pulled back again so that no more Bitcoins or Dollars are bought than was requested. Doing this, UBE would risk the order being filled at two exchanges simultaneously, resulting in more Bitcoins or Dollars being bought than the user asked for. Of course this is unacceptable for a single user, but I wonder - if there are many users trading on UBE - if this risk could be mitigated because of overlapping buy and sell orders from multiple users. So that if User A, User B, User D and User Z all place a bid of $2.30, and this bid is placed on all the available exchanges; then if the bid is met at two exchanges simultaneously so that more Bitcoins are bought than User A requested, they are just used to fulfill User B's bid.

If a traditional exchange doesn't actively prevent it, it could also potentially bring down the trading fee for the users on UBE, since UBE would reach very high trading volumes on its individual exchange accounts that it uses to execute orders.

Are anyone actively developing this? As far as I can see it has a lot of advantages. Of course, UBE would have to charge a fee for this service (on top of the individual exchange fees), but maybe people would find that it's worth it (providing the fee is small enough) since they can get a better price?

The obvious candidate for implementing this, in my opinion, would be Bitinstant.com. They already have exchange accounts on all the major exchanges, but only use this to enable instantaneous movement of funds between them. If they created a trading back-end that would automatically find the best prices on all the exchanges they use, and created a nice web interface for trading, this could become reality without as much hassle as a new company starting up and getting all the required licenses, acquiring capital etc.

[1713864786]

1713864786

[1713864786]

1713864786

[1713864786]

1713864786

[DeathAndTaxes]

The largest issue is the massive amount of capital you would need to provide liquidity.  Even if you had the capital (say $2M USD) you need to earn a return on that capital.  That return would need to be made by fees.  Fees that are on top of exchange fees AND are sufficient to warrant a reasonable return on a large amount of capital.

The slow beyond belief speed on fiat money transfers is what makes a huge amount of capital necessary.

[runeks]

I agree, that would be the biggest issue. That's why I suggested Bitinstant.com, which - as far as I can tell - already has this set up. They have funded accounts at many exchanges, and offer instantaneous transfers between them. Of course, they would probably need even more money if they want to be used to execute all available Bitcoin orders.

Maybe they would need to put a limit on orders to begin with. The largest one-day volume on Mt. Gox was just below 400K BTC. If they needed to have this amount of cash in all their exchange accounts (in order to be able to execute this order on either of them) they would need (number of enabled exchanges)x400K. That's a lot!
But then again, I don't think this would be necessary. The largest one-day volume on Tradehill is only around 20K BTC. So it seems like having 400K available on each exchange wouldn't be necessary.

In any case they can just start small, and only allow orders of 100 BTC or less, as far as I can tell. And they might be forced to close trading on specific exchanges if funds are spent on an account faster than they can refill it.

I think especially the smaller exchanges would love this. Easy competition with Mt. Gox.

[DeathAndTaxes]

I agree, that would be the biggest issue. That's why I suggested Bitinstant.com, which - as far as I can tell - already has this set up. They have funded accounts at many exchanges, and offer instantaneous transfers between them. Of course, they would probably need even more money if they want to be used to execute all available Bitcoin orders.

Maybe they would need to put a limit on orders to begin with. The largest one-day volume on Mt. Gox was just below 400K BTC. If they needed to have this amount of cash in all their exchange accounts (in order to be able to execute this order on either of them) they would need (number of enabled exchanges)x400K. That's a lot!

Actually it would be worse.  To cashout fiat from one exchange and then deposit it into another exchange can take 5-10 days roundtrip.  Thus to handle a worse case scenario where fiat funds are flowing one way you would need 5 to 10 day reserve at each exchange.  In your hypothetical example to cover 400K trade at any of 4 exchanges w/ 10 day delay in rebalancing the fiat would require $16M in cash reserves.

Ironically the worse the exchanges are at moving fiat (in terms of delays, daily limits, etc) the MORE capital you would need to acts as an intermediary. 

Still it is a good idea even if it has a relatively low initial daily caps.

[btc_artist]

Would there be any motivation on the part of the current existing exchanges to facilitate this new exchange?  Perhaps by lowering fees and/or providing easier ways to move funds around.  By my way of thinking, existing exchanges (even the market leader) would win by this, since it would drive more business their way (if they have the best price)?

[DeathAndTaxes]

Would there be any motivation on the part of the current existing exchanges to facilitate this new exchange?  Perhaps by lowering fees and/or providing easier ways to move funds around.  By my way of thinking, existing exchanges (even the market leader) would win by this, since it would drive more business their way (if they have the best price)?

The smaller exchanges would benefit but MtGox would "lose".  It would make their "biggest & most liquid status" less important.  It would be in the best interest of every other exchange to assist though. 

[btc_artist]

Would there be any motivation on the part of the current existing exchanges to facilitate this new exchange?  Perhaps by lowering fees and/or providing easier ways to move funds around.  By my way of thinking, existing exchanges (even the market leader) would win by this, since it would drive more business their way (if they have the best price)?

The smaller exchanges would benefit but MtGox would "lose".  It would make their "biggest & most liquid status" less important.  It would be in the best interest of every other exchange to assist though. 

That makes a lot of sense.  If all the other exchanges teamed up, it would be a good thing for them and for the end customer.  The exchanges (less MtGox) could all agree to lower their fee to something that's really low (0.001%) and assist in keeping funds balanced between exchanges.  In fact this new unified exchange could be a joint project created by all the "other" exchanges and partially owned by them.

[runeks]

Actually it would be worse.  To cashout fiat from one exchange and then deposit it into another exchange can take 5-10 days roundtrip.  Thus to handle a worse case scenario where fiat funds are flowing one way you would need 5 to 10 day reserve at each exchange.  In your hypothetical example to cover 400K trade at any of 4 exchanges w/ 10 day delay in rebalancing the fiat would require $16M in cash reserves.

Ironically the worse the exchanges are at moving fiat (in terms of delays, daily limits, etc) the MORE capital you would need to acts as an intermediary.  

Still it is a good idea even if it has a relatively low initial daily caps.

Good point.
Seems to me that the longer it takes to move traditional currency out of or into an exchange, the larger the fees associated with trading on this exchange would have to be to accommodate for the higher reserve - and thus higher risk - necessary to stay liquid in this exchange. This would create incentive for the participating exchanges to reduce the time it takes to deposit to or withdraw traditional currency from them. Since the fees would be included in UBE's calculations of where to place an order, exchanges would be incentivized to speed up deposit and withdrawal times in order to reduce fees, and make it more likely that a trade is placed with them.

I just had another idea. What about, instead of finding some giant risk-willing investor, we make it crowd-funded? UBE could offer the Bitcoin community that they deposit any amount of Dollars at an exchange, hand over the money to UBE using a redeemable code, and get a 1% interest rate (or whatever is possible) on this deposit. The Dollars would be used as reserves, and the interests paid by UBE would be funded by the fees paid by its customers. That way we spread out the risk between more parties. So instead of one giant investor risking to lose millions of dollars, many people would risk losing whatever comparably small amount they have invested (in return for an interest on their investment).
So I would go to UBE's website, look up where it needs funding, deposit - for example - $100 to my Mt. Gox account, create a redeemable code for this, and send it to UBE. They would be able to use this money as a reserve, and would pay out interests in either Bitcoins (to some specified address) or Dollars (by sending me redeemable Mt. Gox codes for the USD interests every x days). Of course this requires than an exchange enables moving traditional currency easily between accounts on that exchange.

[btc_artist]

I just had another idea. What about, instead of finding some giant risk-willing investor, we make it crowd-funded? UBE could offer the Bitcoin community that they deposit any amount of Dollars at an exchange, hand over the money to UBE using a redeemable code, and get a 1% interest rate (or whatever is possible) on this deposit. The Dollars would be used as reserves, and the interests paid by UBE would be funded by the fees paid by its customers. That way we spread out the risk between more parties. So instead of one giant investor risking to lose millions of dollars, many people would risk losing whatever comparably small amount they have invested (in return for an interest on their investment).
So I would go to UBE's website, look up where it needs funding, deposit - for example - $100 to my Mt. Gox account, create a redeemable code for this, and send it to UBE. They would be able to use this money as a reserve, and would pay out interests in either Bitcoins (to some specified address) or Dollars (by sending me redeemable Mt. Gox codes for the USD interests every x days). Of course this requires than an exchange enables moving traditional currency easily between accounts on that exchange.

To build on that idea, create UBE as a coop. People deposit what they want to be used as reserves, and instead of paying 1% interest, pay a percentage of the income from fees based on the amount they have deposited.  First pay operating expenses from the fees (including salaries, where needed), and divvy up the remaining fee income to the reserve depositors.

[DeathAndTaxes]

To build on that idea, create UBE as a coop. People deposit what they want to be used as reserves, and instead of paying 1% interest, pay a percentage of the income from fees based on the amount they have deposited.  First pay operating expenses from the fees (including salaries, where needed), and divvy up the remaining fee income to the reserve depositors.

That would be both ambitious and interesting.

[btc_artist]

It would be a gargantuan task to create it and get enough people on board for critical mass, but if we did it, it just might work.

[runeks]

Yes, btc_artist, that would be a better idea. Less risk for UBE that way. No guarantees. I do think that whoever gets UBE up and running would have to deposit some money him- or herself into the project. Few people will want to risk their money on providing liquidity for a company that loses nothing themselves by getting compromised. UBE has to have something at stake too so that potential investors know that UBE has an incentive to keep the site secured. The idea should be that if the site gets compromised, and the investor's money are stolen, there would be no guarantees. That's why they're getting a cut of the fees.
I imagine this would also lead potential investors to test the UBE site for security vulnerabilities, and this would be positive as it would create an incentive for the "good people" to find security holes, and thus improve site security, instead of just the "bad people" who want to run away with the money. They would have interest in finding any vulnerabilities and reporting them to the owner of UBE so he could get them fixed.

[Stn]

runeks, this is kind what you were talking about but without massive investment in different currencies: www.centraw.com

[runeks]

As far as I can see that's not what I have suggested in this post. But it certainly looks interesting! This seems to be a site that matches two persons, one who wants to deposit traditional currency (and buy Bitcoins), and another who wants to withdraw traditional currency (and sell Bitcoins).

This thread is basically about a single front-end to all the Bitcoin exchanges out there, where the large reserves needed to accomplish this are provided by the community, in exchange for a cut in the fees.

[notme]

You could just use bitinstant API to move funds between exchanges.  UBE is a great idea, but the biggest problem is not funding, but building it.  If you build it funding the buffers will be easy.

[Stn]

I don't think that accumulating large amounts in someone's hand will be great idea. BTC is not a problem of course, but in other currencies (even electronic) transfer fee may compete with that profit in exchange fee cut.

Also large amount can be vulnerable for attacks. Even if the holder is honest person or group and won't run away, there are hackers, official ID holders and other such kind predators around.

[markm]

Some of this sounds like what I would like to do with my Open Transactions server. It is currently just in alpha testing, see https://bitcointalk.org/index.php?topic=53329.0

However, I would prefer not to deal with fiat.

But that is where the existing exchanges come in. They deal with fiat currencies.

They also deal with all the tiny little "I want to buy 0.01 BTC" kind of stuff, which I would also prefer not to bother with.

I have set up my currency contracts so far to treat coins as whole numbers, and I hope that I will not have to allow decimals since for example if you want to sell some coin or other for 2.99 of some other coin you could offer 100 of the first coin for 299 of the second coin on the units-of-100-coins market of the first coin. I am hoping that using whole numbers might help cut down on the number of frivolously tiny trades.

Open Transactions uses quite secure communications, but that makes it not all that handy for web-based use, as the kind of people who insist on using a website tend to also insist on using pathetically easy to guess passwords and stuff like that. So basically this would not be targeting the casual user who deals in such trivial amounts that they basically do not bother with much security. Such people are just going to give the service a bad name anyway by constantly managing to get ripped off, hacked, trojan'd etc etc anyway. So we'd hopefully be targetting serious people who want the good security.

Probably instead of even having MtGox USD at all, I'd only have MtGox BTC, and let people do the conversion to fiat over at the fiat-to-coins exchange. I'd try to focus on *coins and on stocks, shares, bonds type of things.

By not dealing in fiat I would presumably be able to have much better liquidity across the various exchanges that do handle fiat.

Also, maybe I could have tokens that represent BTC actually held at third party exchanges, so one could trade that without having to actually move the BTC from exchange to exchange. The added risk though compared to tokens representing BTC I actually have in my own secure wallet might make the tokens representing coins held at third party exchanges trade at a different price than those representing coins I have direct control of?

-MarkM-

[offshoregenius77]

It would definitely be a solid idea. You would need to raise a massive amount of capital though.