Decentralized Exchanges vs Centralized Exchanges

[Akado]

One positive thing I forgot about UIAs is that people can use arbitrage to trade

[Samupaha]

This is how I see it all will go:

Exchanges who share their orderbooks will become mostly gateways. They don't need to offer an exchange service for their customers, because that's what Bitshares does. They will be just a UI for BEX – preferably very good UI so that they can get lots of customers and referral income.

Normal exchange works like this: it offers the exchange service and customers can trade USD-IOU and BTC-IOU in there.

When a normal exchange starts to use BEX as a trading engine, they will still issue their own IOUs and let customers trade with those. I don't see a problem with that. There will be many gateways and many different IOUs for same currency, like BTC or USD. Some of those, like USD, will not be allowed to trade against all other assets for regulatory reasons. But others, like BTC, will be.

So there will lots of trading pairs that will create big markets. This is good from Bitshares point of view because it will create much more trading than only bitBTC alone would create.

BTC-IOUs from different gateways will trade 1:1 for most of the time. But if there is some suspicions of insolvency, IOUs from that particular exchange will become cheaper and that will provide valuable information for all market participants. IOUs are not backed with nothing, they are backed with reputation.

Stan has a good point on risks and time. And I'd like to emphasize that for most ordinary people, gateway-IOUs will be only a tool to get in or out the system. We can let more advanced traders to handle the trading with all the different IOUs. I think they might like if there were lots of different trading pairs and not only a few.

[Stan]

I see, that makes sense and you've placed the question from a point I wasn't seeing.

Risk is proportional to the amount of time you hold a certain asset.

Trustless assets tend to be kept as a savings account, for hodlers while UIAs (ie OPENBTC) can be used by the exchanges joining the network as they will mostly be used for trades.

Makes sense. However, the less risk one is exposed to, the better, we still should strive to find a better way, to maximize the usage and adoption of trustless assets (even though their risk is always present too, it's just small).

What I mean is, that's always an argument against us. Of course there's no perfect solution, that's why we have so many different projects competing, we're just testing the waters, experimenting, trying to get better, but it can be considered as a "weakness" or an argument against BitShares.

Assuming other exchanges join OpenLedger.
Assuming they use OPENassets

Can the risk of manipulation be decreased with the use of multisig? Could, for example, the permissions of the asset b extended to multiple trustworthy people like the committee? If that is possible, in a way couldn't we be compared to Nubits and the Fed because a selected group of people have the power over an asset? (Note that we can vote in and out the members who would hold the power so there's always room for people to choose who they trust with that power)

I think most of the potential clients out there are not yet aware of most of these advanced issues and would be happy using assets that are controlled by mechanisms with varying degrees of trustworthiness engineering without understanding the fine points.

Certainly one way to compete is to make improvements in this area and then raise awareness of maturing users about what differentiates each product and/or service.

But initially, I think folks will just be amazed with the simple concept that their funds are far safer than they currently are in the fiat world.  After that quantum leap in insight, then they can get educated into the finer differences between various asset management strategies.

(And the more they learn about those fine points, the more they will appreciate how far ahead the BitShares platform is beyond The World Financial System that they had been using.)

[Akado]

I see, that makes sense and you've placed the question from a point I wasn't seeing.

Risk is proportional to the amount of time you hold a certain asset.

Trustless assets tend to be kept as a savings account, for hodlers while UIAs (ie OPENBTC) can be used by the exchanges joining the network as they will mostly be used for trades.

Makes sense. However, the less risk one is exposed to, the better, we still should strive to find a better way, to maximize the usage and adoption of trustless assets (even though their risk is always present too, it's just small).

What I mean is, that's always an argument against us. Of course there's no perfect solution, that's why we have so many different projects competing, we're just testing the waters, experimenting, trying to get better, but it can be considered as a "weakness" or an argument against BitShares.

Assuming other exchanges join OpenLedger.
Assuming they use OPENassets

Can the risk of manipulation be decreased with the use of multisig? Could, for example, the permissions of the asset b extended to multiple trustworthy people like the committee? If that is possible, in a way couldn't we be compared to Nubits and the Fed because a selected group of people have the power over an asset? (Note that we can vote in and out the members who would hold the power so there's always room for people to choose who they trust with that power)

[Stan]

The problem still remains, shared order books are an awesome concept, it can be something really big, however I just can't get past the problems with the collateral needed to create the assets. I would like to know people's opinions on this and brainstorm a little.

How other exchanges get the assets that users trade? From the users, when they do deposits. My guess is that other exchanges that will share their books with Bitshares are going to focus on profiting via gateway and referral businesses. They just try to get as much users as they can and let them trade whatever they want.

BitBTC isn't the real bitcoin. It's a derivative that is useful only in Bitshares exchange and nowhere else. That's why I don't think we can presume that all exchanges are going to do 1:1 rate from real BTC to bitBTC. Instead they will let customers trade and free markets will decide the exchange rate.

Yes, I only choosed bitBTC since it's backed by collateral and no entity owns it, meaning it's the safest. Our only option is OPENBTC since they will run on OpenLedger. However that's not backed by anything and can be created at will.

We trust Ronny or whoever holds the keys that they will only issue more OPENBTC as long as user's deposit BTC on the exchange right? Well, but can we be sure they do it and don't make any mistakes like printing more OPENBTC than needed?

1st we need all exchanges to share the same assets, otherwise there won't be any shared order book and I think most people here are forgetting about that.
Then we have the problem of finding that asset. bitBTC and we have the collateral problem? Or OPENBTC (for example) since it requires no collateral? How can we be sure there aren't any more OPENBTC around than truly should be?

My problem with that is that it defeats the purpose of decentralization. The only way to be sure one asset is created in a safe and decentralized way is by creating bitBTC, bitUSD, etc. Not by UIA.


Of course ultimately it's up to the markets to decide. But apparently we can't have the best of both. We either have shared order books or we have safe, trustless assets... The way our system is designed is not for both.

Sure anyone can create their own UIA and that's useful, but we should focus on trustless assets right? Otherwise there is no decentralization. It defeats it's own purpose. I think people aren't really worried about that. bitAssets were supposed to be the "thing", but they're slowly being replaced by UIA because service providers can't keep up with the collateral. That's something we, imo, should worry about. That will cause us to use 3rd party assets, meaning there's really no difference between how the current model worls. If we do that we can't claim our dex is trustless.

Anyone can choose if they prefer a trustless asset or a 3rd party asset of course. But in the end trustless assets won't be the ones being promoted. It will be the UIAs from each exchange. A UIA can be manipulated so, there's no point in having it from a costumer's pov.

I think people haven't really thought this through. In the end, the way the system is designed, it's not favourable to trustless assets if we consider this specific example of exchanges joining in. Which, although specific, should be one of our targets but that's just my opinion. And I don't see anyone else concerned about this.

I think there is a place for trustless and non-trustless assets to coexist side by side.

Risk is the integral of time spent holding a trusted asset.  If you are going to trust a third party with a little bit of money for the few seconds or minutes it takes to pass through whatever centralized service they are providing, then that is quite tolerable and not very different from how the whole world economy works.

If you are going to hold a large amount of value for a long period of time, then that's where a trustless asset is needed most. 

One might think of it like you are using trusted for checking and trustless for savings.  People usually have both and expect them each to work a bit differently.

The BitShares Exchange Network lets you move between trustless and trusted assets in a seamless fashion.

[Akado]

The problem still remains, shared order books are an awesome concept, it can be something really big, however I just can't get past the problems with the collateral needed to create the assets. I would like to know people's opinions on this and brainstorm a little.

How other exchanges get the assets that users trade? From the users, when they do deposits. My guess is that other exchanges that will share their books with Bitshares are going to focus on profiting via gateway and referral businesses. They just try to get as much users as they can and let them trade whatever they want.

BitBTC isn't the real bitcoin. It's a derivative that is useful only in Bitshares exchange and nowhere else. That's why I don't think we can presume that all exchanges are going to do 1:1 rate from real BTC to bitBTC. Instead they will let customers trade and free markets will decide the exchange rate.

Yes, I only choosed bitBTC since it's backed by collateral and no entity owns it, meaning it's the safest. Our only option is OPENBTC since they will run on OpenLedger. However that's not backed by anything and can be created at will.

We trust Ronny or whoever holds the keys that they will only issue more OPENBTC as long as user's deposit BTC on the exchange right? Well, but can we be sure they do it and don't make any mistakes like printing more OPENBTC than needed?

1st we need all exchanges to share the same assets, otherwise there won't be any shared order book and I think most people here are forgetting about that.
Then we have the problem of finding that asset. bitBTC and we have the collateral problem? Or OPENBTC (for example) since it requires no collateral? How can we be sure there aren't any more OPENBTC around than truly should be?

My problem with that is that it defeats the purpose of decentralization. The only way to be sure one asset is created in a safe and decentralized way is by creating bitBTC, bitUSD, etc. Not by UIA.


Of course ultimately it's up to the markets to decide. But apparently we can't have the best of both. We either have shared order books or we have safe, trustless assets... The way our system is designed is not for both.

Sure anyone can create their own UIA and that's useful, but we should focus on trustless assets right? Otherwise there is no decentralization. It defeats it's own purpose. I think people aren't really worried about that. bitAssets were supposed to be the "thing", but they're slowly being replaced by UIA because service providers can't keep up with the collateral. That's something we, imo, should worry about. That will cause us to use 3rd party assets, meaning there's really no difference between how the current model worls. If we do that we can't claim our dex is trustless.

Anyone can choose if they prefer a trustless asset or a 3rd party asset of course. But in the end trustless assets won't be the ones being promoted. It will be the UIAs from each exchange. A UIA can be manipulated so, there's no point in having it from a costumer's pov.

I think people haven't really thought this through. In the end, the way the system is designed, it's not favourable to trustless assets if we consider this specific example of exchanges joining in. Which, although specific, should be one of our targets but that's just my opinion. And I don't see anyone else concerned about this.

[Samupaha]

The problem still remains, shared order books are an awesome concept, it can be something really big, however I just can't get past the problems with the collateral needed to create the assets. I would like to know people's opinions on this and brainstorm a little.

How other exchanges get the assets that users trade? From the users, when they do deposits. My guess is that other exchanges that will share their books with Bitshares are going to focus on profiting via gateway and referral businesses. They just try to get as much users as they can and let them trade whatever they want.

BitBTC isn't the real bitcoin. It's a derivative that is useful only in Bitshares exchange and nowhere else. That's why I don't think we can presume that all exchanges are going to do 1:1 rate from real BTC to bitBTC. Instead they will let customers trade and free markets will decide the exchange rate.

[Akado]

Good points Akado!

My suggestion is that we emphasize less the decentralization part of our exchange and instead call it what it really is: derivatives exchange. Let bitcoin be just a tool to get in and out the system via gateway and not focus on it as an important trade object.

We can have hundreds of different markets that other exchanges don't have. Let's forget the bitcoin and concentrate on getting new customers to come to our exchange and trade and use our unique products.

Well I only used Bitcoin as an example. It can be Bitcoin or any other asset.

The problem still remains, shared order books are an awesome concept, it can be something really big, however I just can't get past the problems with the collateral needed to create the assets. I would like to know people's opinions on this and brainstorm a little.

Also any other pros and cons of decentralized vs centralized exchanges.

[Samupaha]

Good points Akado!

My suggestion is that we emphasize less the decentralization part of our exchange and instead call it what it really is: derivatives exchange. Let bitcoin be just a tool to get in and out the system via gateway and not focus on it as an important trade object.

We can have hundreds of different markets that other exchanges don't have. Let's forget the bitcoin and concentrate on getting new customers to come to our exchange and trade and use our unique products.

[Akado]

Yes, first mover advantage is really important, still, if a smaller exchange comes out with something innovative, develops it and then a bigger one does the same thing after, the bigger one is still the one with the advantage because it has more exposure. But I guess that's just how the world works.

One other thing that I wanted to know and you replied are the factors that can differentiate each exchange while they all have the same thing and potentially even the same features. So they're the following:

- Costumer Support (Help provided to the clients)
- Interface (pleasing, efficient, nice to use)
- On/Off ramps (spreads)
- First Mover Advantage (this would be one more incentive to keep exchanges innovating)

Does anyone know of something more? In resume, it's the costumer experience. The one who can provide the best experience wins, but I would like to separate that into several parts as to know what the clients will evaluate and base their opinion on! I think the Exchange's own community matters too so I'll add that

- Community (useful members, trollbox)

Anything more?

@Stan You contradicted yourself I believe. All exchanges can have shared marketing since they are based on the same platform. However to achieve shared order books in order to increase market depth and liquidity - and this is just the most awesome thing - you can't have exchanges competing with UIAs to see which one offers to lower spread.

What they can offer is a bigger variety of offers. For example,  exchange A offers the top 10 cryptos. Exchange B offers the top 4 but with multiple other assets from their own businesses that are offering nice returns. Then it's up to the costumer to decide which one to use.

But if we have exchange A with A.BTC and exchange B with B.BTC competing to offer the smallest spread on BTC price, they won't be having shared order books. I guess that, itself is another factor client's might take into consideration.

One other question is: What incentives does an exchange have to offer bitBTC, with shared order books, than issuing their own UIAs? For bitBTC they need to have a huge amount of collateral. That may prove itself challenging for an exchange. Imagine they have 1000 BTC profit. They can create what, 500 BTC? their btc supply will be cut in half. If one costumer wants to deposit BTC on an exchange running on OpenLedger, how can he do that with bitBTC? If no more bitBTCs are created, there's not enough supply.

The exchange can't just create one bitBTC per BTC sent there. So how will one exchange handle that? I think we should have the answers in order to promote shared order books and be ready to help if any wants to join.

If a costumer deposits 100 btc, the exchange can't create 100 bitBTC. It can only create 30-50 right? How will that be handled? Client's won't ever accept their BTC to be temporarily not available..... What's the solution? An exchange already has to have a large pool of bitBTC to keep up with deposits right? Then:

- How can we get an exchange to create such an amount of bitBTC?
- What if costumer's deposits are more than the exchange's reserves of bitBTC?

Could we get around this with some marketing move? Here in Portugal there was an awesome service of movies and tv series streaming, however since so many wanted to join they restricted access, only the ones joining during the first years could access it. Could we do the same? For starters I doubt most would accept it, however,

- That would help us create a brand and distinguish ourselves as a "top" exchange
- We're the safest exchange and have the best services so only a select few could join
- If someone wants to join they would have to pay a fee

That seems to go against the concept of crypto and adoption itself.. It could be both a huge marketing move or a huge marketing failure. Not to mention with time other decentralized exchanges would appear and as the market gets competitive, people will always go for the option which is free and does not restrict it's users. We would need to be at the top of the exchange world to ever consider doing that. However I don't see any other option atm...

Exchanges, even if they don't charge clients, would need to restrict deposits at some point in time. And success could be the cause of this. The more successful it is, more costumers want to join in, which means more deposits, but how can an exchange keep up with all the deposits if each deposit require a 150%-200% collateral?

Can any business in the world even be profitable enough so the profits can compensate for that? Even if that happened, what would be left for the exchange owners? I just don't see an exchange placing 500k in collateral to create bitAssets with the constant need to increase the amount in collateral to have bitBTC and bitAssets. That doesn't seem like an incentive at all. And without that, we cant have shared order books which would be one of the biggest things I've seen.

What will they do then? All use OPENBTC? That's not backed by anything. There's a risk there that the owner of that asset messes up the market. Meaning we're not that trustless.

[Stan]

Hey guys I would like you to share and discuss the pros and cons of both so that I could organize that info to complement with my list of exchange issues

Decentralized:
- Safety, can't be hacked / funds stolen
- Transparency no shady activities from the exchanges as everyone can audit them through the blockchain
- Shared Order Books which can increase market depth and liquidity
- Can act as a Network of Exchanges
- Potential for having the best features. Since they can be implemented at the blockchain level, if Bond Markets are allowed on chain, every Exchange running on that chain has always the best tech. It only takes someone to develop it.

Centralized:
- You can bridge crypto/fiat and vice versa


Still, you can place a centralized service running on top of the DEX. Can you give me more pros and cons of both and develop more on this please?
What about competition? How can they compete? As I mentioned imagine Prediction/Bond Markets are implemented, every exchange has that same feature. This means exchanges can't differentiate from each other and stay more competitive, could they? If so, how? I think only through their gateway to fiat really.... What more can 1 exchange offer that other can't, if features are implemented at blockchain level? On the other side, if a feature isn't on chain and is done by a 3rd party, the risk to fail and compromise the costumer is higher. How can we solve this?
Could one exchange develop one feature imagine through FBAs and then sell it to other exchanges for profit later on? But how could that original exchange keep the tech only for itself? How would the access to that feature be restricted?

Imagine we have Exchange A and B running on top of OpenLedger. One funds Prediction Markets so it's only their right to profit from it. How does it restrict the usage for their own users? Can that be done if each exchange has a different wallet for example? While still being part of OpenLedger?

So exchange A would have its costum wallet with Prediction Markets built in the GUI so users could access it. Exchange B would also have its own custom wallet but without Prediction Markets. Still, what keeps exchange B from paying to a 3rd party for PM integration on their own wallet? The problem here is if exchange A was the original one funding the development of PM technology on chain first. Exchange B would take advantage paying only a fraction of the cost. Can that be done via FBAs? If so, how?

Any other pros/cons you know of?

Would really appreciate some help with this please

These are AWESOME questions!  You have really been thinking!   I love it.

Each member of the network gets the network fees from the customers they sign up regardless of whose services they use.  Each member gets the service-specific revenues from everybody's customers who use that service.  Shared marketing and shared market depth benefit everybody's products and services.

Each member competes to provide the best interface, customer support, and auxiliary services.  Who's on/off ramps are most versatile and cost effective?  Who's UIA assets are most trusted with the best spreads?  There are a thousand ways to compete while sharing a common infrastructure.

(Why do you think car dealers and restaurants always locate on the same street?)

Each exchange (or other enterprise) presumably has its own independent customer demographic where there is little to no initial competition.  So, in the critical bootstrap phase, its all benefit and no drawback to share market depth and liquidity.

After that, competition is all about what you build on top of the network and how well you earn your user's loyalty.

As for having someone clone your FBA, this is a problem with open source code that we need to resolve at the community level.  Do we want to grant exclusive licenses?  How can we attract investors if they aren't guaranteed the right to profit from what they funded?   How does that work in the altcoin world?  Usually, its about first mover advantage and network effect.

[Akado]

Hey guys I would like you to share and discuss the pros and cons of both so that I could organize that info to complement with my list of exchange issues

Decentralized:
- Safety, can't be hacked / funds stolen
- Transparency no shady activities from the exchanges as everyone can audit them through the blockchain
- Shared Order Books which can increase market depth and liquidity
- Can act as a Network of Exchanges
- Potential for having the best features. Since they can be implemented at the blockchain level, if Bond Markets are allowed on chain, every Exchange running on that chain has always the best tech. It only takes someone to develop it.

Centralized:
- You can bridge crypto/fiat and vice versa


Still, you can place a centralized service running on top of the DEX. Can you give me more pros and cons of both and develop more on this please?
What about competition? How can they compete? As I mentioned imagine Prediction/Bond Markets are implemented, every exchange has that same feature. This means exchanges can't differentiate from each other and stay more competitive, could they? If so, how? I think only through their gateway to fiat really.... What more can 1 exchange offer that other can't, if features are implemented at blockchain level? On the other side, if a feature isn't on chain and is done by a 3rd party, the risk to fail and compromise the costumer is higher. How can we solve this?
Could one exchange develop one feature imagine through FBAs and then sell it to other exchanges for profit later on? But how could that original exchange keep the tech only for itself? How would the access to that feature be restricted?

Imagine we have Exchange A and B running on top of OpenLedger. One funds Prediction Markets so it's only their right to profit from it. How does it restrict the usage for their own users? Can that be done if each exchange has a different wallet for example? While still being part of OpenLedger?

So exchange A would have its costum wallet with Prediction Markets built in the GUI so users could access it. Exchange B would also have its own custom wallet but without Prediction Markets. Still, what keeps exchange B from paying to a 3rd party for PM integration on their own wallet? The problem here is if exchange A was the original one funding the development of PM technology on chain first. Exchange B would take advantage paying only a fraction of the cost. Can that be done via FBAs? If so, how?

Any other pros/cons you know of?

Would really appreciate some help with this please