Topic: Any work being done on decentralized fiat exchanges? (Read 497 times)

Yep. I think you got it down. I think the trigger would be something like "if volume on the decentralized exchange is higher than the volume of the largest exchange (or all of them combined), then set price to the swap price on the dex". I think if the dex is one of the largest exchanges, the difficulty to manipulate the underlying value would be pretty high.

And I do believe the address that is filled with BTC will need to play a large role in maintaining stability but I'm having a hard time figuring out how to do it programatically without needing to go through Rootstock or ETH. Perhaps some sort of game theory which incentivizes the org in charge of the receiving address to maintain stability (like they charge a percentage above the price, and buy back at a percentage that it drops...thus constantly requiring the new issuance of tokens at the current price.

Ah, ok, I misunderstood. I thought the issuer would sell his BTC on this address for 1 sat/1 cent, which would be exactly like Tether


So let me resume the model, as I understand it now:
- everybody has the right to issue currency tokens (I'll call them CFIAT for now) based on the current price, taken from the "namecoin list of exchanges", according to the satoshis he sent to the issuance address. The software would have then to prove that  the issuer has issued the quantity of currency corresponding to the valid price of the moment of issuance - otherwise his tokens wouldn't be valid.
- In parallel, a decentralized exchange is built up, where people can trade BTC/CFIAT via LN atomic swaps, and CFIAT/FIAT pairs via a Bisq-like protocol using LN.
- Once the decentralized exchange has enough volume, a "trigger" determines that the "list of exchanges" is replaced by a decentralized price finding mechanism.

I think I understand now: You hope that when the price is "liberated", the token is already so popular that there is practically no volatility anymore with respect to fiat.

For example, there could be a market like I proposed here for Bitcoin, where merchants could "guarantee" a price in "cryptofiat" for a certain time e.g. 24 hours or a week. Knowing that, if there are price swings to the downside (which is the biggest risk), arbitrators could profit buying the products with "cheap coins". And there could be also emerge exchangers which always would back the token with 1 CFIATUSD=1USD for example.

Still there is the problem that someone that wants to manipulate the price and has enough Bitcoins or cryptofiat tokens can do it. Not even the elaborated "stablecoin" models like BitUSD or Dai are entirely safe from this kind of attack. But it may work if there are enough backing mechanisms in place like those I mentioned in the last paragraph.

Looks definitively like an interesting concept, I have to think about other possible problems.

There's not so much an issuer of the cryptofiat as basically anyone sending BTC to that single address is now able to "issue" tokens from their own address (up to the amount they "purchased"). The address could even be a burn address like 1XXXXXXXX... But I have been kicking around the idea that the address is used to fund further development or even buy back cryptofiat if the market deems it necessary. Not a single person but likely a decentralized organization of some sort with multi-sig access to the address.

The control of that address could disappear (keys lost, people controlling it die off, etc.) and the system would still continue on with no disruption.


I use bisq as reference but it would likely be a modified version so a user does not necessarily need to download any software (make it web based). While still being able to download software and avoid the middle man.


This part is a bit unhashed but I believe that it will need to start small and grow (likely starting in a small country with few or no exchanges). It will start as a single server but anyone who downloads the software and wants to run their own server (exchange) can also do so. The exchange could charge fees or advertise or whatever they need to make money which allows them to enhance the user experience and advertise to draw more people in. With more volume on the exchanges they're all using the p2p in the background. It's just that the user doesn't know (or shouldn't even care).

The point of the namecoin list of exchanges is so that initially the price can be set to a specific formula (ie. take the volume from Bitstamp, Bitfinex, GDAX, etc. combine and divide to get the price). The price will need to be known by everyone so that the cryptofiat can be the price of the underlying currency while still using satoshis (ie. when someone sends BTC to the initial address at $7k/BTC and 1 satoshi = $1, then everyone in the network knows that the person can issue 7k satoshis and they can ignore any further satoshis sent from that address by consensus.

Since there will be such low volume initially, the price on the decentralized exchange will be too easily manipulated.

But you can't just have a single issuer deciding the price to be used. So initially it will be centralized, then grow more decentralized over time. Then at a trigger point the price is based on the price on the decentralized exchange (due to high enough volume). I figure the "issuers" should be given the ability to determine the price formula that is used.


It certainly is like a decentralized Tether. The main problem with Tether is that it's run by a single company. They have(had) a bank account. Disruption of that single company/account can bring it all crumbling down.

The Bitcoin price can (will) fluctuate no problem. When issuer X buys at $7k they get 7k satoshis. If the price jumps to $10k when issuer Y buys then they can issue 10k satoshis.

Trade on other platforms is more than welcome. If it is known that 1 satoshi = $1 then not only can it just be used for exchange, it can be used in commerce. Merchants can start accepting cryptofiat alongside regular fiat. My ultimate dream would be that cryptofiat becomes so much easier to use than regular fiat that nobody uses regular fiat anymore. Banks become useless and eventually when BTC price reaches the point where 1 satoshi = $1, people will now just use their cryptofiat as bitcoins. Essentially, bitcoin has now replaced a government currency.

Trust is certainly essential. I believe that if a single exchange is run like this for a while and reaches high enough volume then it will sort of prove itself. Initially it will be centralized and controlled with the mechanisms built in to grow more decentralized over time to the point where it is fully decentralized. People already trust that when they send money to an exchange that the numbers in their database correlate to their money. A lot less trust than I would give but these exchanges are trading hundreds of millions each day so obviously some people trust these database entries.

Thank you for the detailed description. I have some questions/doubts and remarks about it:

So, if I understand right:
- There is one single issuer of the cryptofiat
- Distribution is only via LN (which makes sense, due to the low amounts of "real" BTC to be expected if 1 sat = 1 unit, e.g. 1 cent), the purchasers having to open a channel to the issuer.

Question: Wouldn't the issuer need a lot of trust for the price of e.g. 1 satoshi = 1 cent to be accepted? From my perspective, the model sounds similar to Tether at first, and he would probably need real fiat (e.g. on a bank account with regular audits ...) to back his cryptofiat, otherwise nobody would buy it.

Bisq would need to become LN-compatible for that to work, but I guess you know that.

I don't understand the concept of the "namecoin list of exchanges" and the voting mechanism. Are there multiple (centralized) exchanges involved? I thought there was initially one exchange. What do the users vote, the price or the exchange list?

Unfortunately, I don't think that will work, or I don't understand the mechanism. You are saying that a single purchase of the cryptofiat token with Bitcoin will set the price (potentially forever?)? What is if Bitcoin's price fluctuates heavily, how would the price be "rebalanced"? What if the token becomes traded at other platforms, without the fixed price, and manipulated to the upside or to the downside (something what has happened to the Steem Dollar, for example ...)?

Or are you referring to the "cryptofiat" price measured in fiat? While that would solve the first problem, the second problem (trading at other platforms) would persist. There would still be a "backing" (fiat funds controlled by the issuer) needed to establish trust on the cryptofiat token.

From your explanation, I don't see many differences to Tether - the price is set artificially, and is meant to stay at that price.

I think there is something important missing in your explanation, or something I didn't understand still, so if you want, please elaborate on that.

Yes, I definitely think the Bisq to a cryptofiat would be the key thing to enable more widespread decentralized exchange. The swap from cryptofiat to cryptocurrencies is easier from there.

My proposed solution (and I am currently putting together the pieces to test this out) would be:

-Have a single BTC address from which you can purchase cryptofiat. Initially the price would be set and known throughout the network based on a namecoin list of exchange addresses (more on that in a bit).
-The addresses that purchase the cryptofiat would be known by the network and the amount they purchased could be sent from them in colored coins, with 1 satoshi = 1 unit of the currency, to open a channel on a cryptofiat Lightning Network.
-Bisq can then be used to convert fiat to cryptofiat. User can then open a Lightning channel to receive their colored coins.
-With those colored coins they can easily do an atomic swap for bitcoins, or do multiple micro swaps so that someone can sell just the smallest unit possible adding up to a full purchase rather than needing one person to buy the full amount offered by another person (ie. 50 people can buy from someone offering 1BTC for 7k cryptofiat).

Now back to the cryptofiat price. Initially it would be centralized. It would likely start as a centralized exchange with a website, etc. But using all of the pieces in the background. The user doesn't really even know the difference. Those that buy cryptofiat with bitcoins can start to have some input on the namecoin list of exchanges used (some sort of voting mechanism based on amount purchased and time passed).
Eventually, the goal would be to set the price to the publicly known price of the last trade in the decentralized swap. But initially the volume would be too low to do so. But when that is triggered it becomes truly decentralized and unstoppable.

No tokens needed, ETH not necessary, Rootstock not needed. All done by consensus through Bitcoin. The hope being that a typical user would just log into a website (or open an app), put in how much money they want to send, they get the highest rated Bisq seller for their criteria (bank, amount, location, time, etc.) and they just send their money. They see that they have money "in their account" and can now purchase bitcoins (just like on any other exchange). They purchase it and withdraw their BTC. Same in the other direction.

I had proposed something very similar here in May, but it was only a concept, there is no work done currently on it.

The idea, in a nutshell: Use the Bisq protocol for fiat-to-stablecoin trades and the BarterDEX atomic swap protocol for trade pairs of stablecoin-to-bitcoin/other cryptos.

The reason for the stablecoin integration is that atomic swaps to a "stable" currency would be theoretically possible, so using the exchange to "hedge" against the fluctuations of Bitcoin's prices is possible, without having to use a centralized IOU or exchange. To "get in" or "get out" of the crypto world (cashing out to fiat or buying for fiat) without exchange price risk, Bisq is actually (imo) more suitable than to exchange volatile cryptocurrencies to fiat.

One could use Ethereum-based Dai as an USD-based stablecoin, which seems to have actually most acceptance (from the more "decentralized" types). As Ethereum is supported by BarterDEX, it should be possible to adapt the swap transactions to transfer Dai-to-BTC instead of ETH-to-BTC. The only problem would be the management of "forced settlements", which are always possible in Dai if the exchange rate diverges too much from the target price.

There may be an intermediate step necessary, if the stablecoin used cannot be exchanged directly via an atomic swap. For example, BitUSD seems not to support the BarterDEX protocol, if I'm not wrong. In the BitShares case one could use a combination of the internal BitShares exchange and BarterDEX; this would be however not desirable, so actually Dai may be a better option.

The bank can just have a webpage display a stableUSD address they own for every person wanting to exchange to regular USD

This is pretty simple stuff (maybe they might allow an API for doing this programmatically, but I'm afraid I might lose you if we get into that )

You can BUY stableUSD that way, yes, but you can't sell it back to them like that, you have to do a cryptocurrency transfer to their wallet address.

This way they can also force kyc for each stablecoin tx :-D
but how is this different from actual online banking?

It's simpler just to interchange USD for stableUSD directly on a one-page website, without using any tech inbetween (coloured coins, whatever). The Federal Reserve accepting 1 stableUSD for 1 USD is the most guaranteed peg possible

Then my question still stands: why invent your own cryptocurrency in order to do that?  Its much easier to use a coloured coin.

Right, it's the same model as Tether, but with either the issuing central bank or Wells Fargo etc doing the 1:1 redemption

When coins are in the banking network, they get burnt to nothing, so they don't exist. When they're out in the open, they have to be inside a cryptocurrency to prevent double spends.

Direct convertibility

Either the coins are in the ACH legacy banking network, or they're in p2p stablecoin network. Neither permit double spends. It's simple.

I'm sorry, I don't follow what you're saying. Why doesn't this allow double spending the stablecoin (or double convertibility to use your terminology)?

No the idea would be to use a p2p network protocol to allow permissionless transfer of the stablecoin, but use direct convertibility with the issuer to maintain the peg.

That's not a cryptocurrency you're talking about, its a wire transfer. As you know, cryptocurrencies exist because of the double spend problem - so why go to all the effort of solving it again in your own crypto when you can just piggyback on the most trusted crypto in the world?

So the Czech central bank?




Central banks would be unlikely to take that approach, but in principle there's no technical reason why not. The only reason I have for saying central bank cryptocurrency would solve the stablecoin issue is the simplicity, why use omni layer when they could just redeem TCP/IP routed USD 1:1 directly at federalreserve.com, or through whatever regular commercial bank? 

They don't even need to launch their own cryptocurrency - all they need to do is to create a coloured coin on omni ledger (like USDT is). They'd then have all the power to issue/confiscate coins and trace transactions that they require while piggybacking on bitcoin.

Beneficiary Bank: Fio bank
Beneficiary: CashTan Financial Services s.r.o.
Beneficiary bank address: Fio banka, a.s., V Celnici 1028/10, 117 21 Praha 1
Account No.: 2601191382