I think fees is about the stupidest reason to avoid centralized exchanges.

EVERYTHING has cost.  Bitcoin itself has cost and fees pay that cost.

The things I dislike about centralized exchanges are:
* funds controlled by 3rd party (I would like to see an m of n type system where I have control of my BTC at all times)
* complete lack of accountability (MtGox deciding it is their job to be the Bitcoin Police)
* lack of transparency (does MtGox has $1 USD for $1 USD and 1 BTC for every 1 BTC in client accounts ? )

So far for more users those issues are dwarfed by the benefits of liquidity and narrow spreads.

Competition between the exchanges is what protects the bitcoin traders. If liquidity improves, you'll see even more competition. Right now the market for Bitcoins is small. But still there is decent competition. Look at the BTC/USD market on Bitcoincharts. There are 14 exchanges listed with various fee structures. Some you pay, some are free. There is even 1, bitfloor, that pays you for orders that add liquidity to their order book (you pay to remove liquidity). Competition is the key and I hope there will continue to be competition between the bitcoin exchanges.

I've said this before, but if we had a way to embed an accurate and timestamped Bitcoin trading price into each new block created, it would go a long way toward creating decentralized trading apps. If there was only a way to calculate the average trading price based upon each and every Bitcoin transaction itself, then Bitcoin could have a universal value and could then be calibrated like (or to) the SDR.

This sounds lovely...but a Bitcoin lacks intrinsic value, we cannot eat it, drink it, or live beneath it. How could we ever determine its value?

I believe supply and demand should dictate its value..

There a working model exist exactly what your talking about: www.centraw.com
There is no trust rating implemented but it needed only for reversible systems like Paypal. I personally believe that in perspective reversible payments should became redundant.

Should admit that centralized financial system got one positive aspect. They maintain more or less stable rate of the certain currency. Using special stabilization funds they absorb large supply and demand jumps. It protects local economy.

Supply and demand dictate is good in theory but in practice that is probably the only thing now which prevents Bitcoin from wider adoption. I don't have any ideas know how to solve it. Anyone around does?

Ever heard about King Midas? He figured out that gold does not have intrinsic value either.

My exchange wishlist/ideas:

• Stop loss orders
• Shorting (not a big deal really, but it would be much more efficient if it were available in the exchanges instead of via bitcoinica)
• Live orderbook that is reliable and accurate (Mt.Gox's is manifestly not)
• One-click "acceptance" of an offer in the orderbook.  So I don't have to frig around switching pages, typing numbers in just to say "yes" when someone is bidding at $5 for 3BTC.  I should just be able to click "buy" on a line in the orderbook.
• Overorders allowed.  Mt.Gox does this, but Intersango doesn't.  Without this it's very hard to spend 100% of your balance -- you're always left with a little bit left over.
• Pending orders allowed.  Mt.Gox does this, but Intersango doesn't.  If I were a merchant, I might want to leave a permanent "sell 100,000 BTC" on my account, then give out my Mt.Gox deposit address to customers.  Any coins would then automatically convert to fiat whenever they were received
• Inter-exchange cooperation.  Bitcoinica already kind of does this -- allowing easy code-based transfers between Mt.Gox and Bitcoinica.  It means I can move funds instantly.  Why don't all the exchanges get together and agree to a level of trust between each other; then allow fund transfer?
• When you really get down to it, the exchanges aren't primarily about exchanging bitcoins (perversely), they are the interface to the existing financial system.  Bit-instant has the right idea.  All the exchanges would benefit from being able to use other exchange's bank accounts as their interface.  It would provide a comfortable redundancy for the users too -- if Intersango's account breaks (again), I would know that I could still use Mt.Gox's account to transfer funds out. (I don't hold out a lot of hope for this one)
• A decent buy-sell interface.  I really shouldn't need to get a calculator out to work out amounts when using the exchange.  Mt.Gox has a reasonable implementation: when you type in the USD field it fills in the BTC field automatically, and vice-versa.  It could be better though -- there is a difference between BUY BTC and a SELL USD from a user perspective, and affects how the trade engine should handle slippages
• I like Intersango's liquidity discount on the fees.  It should go further though: there should be profit (not just discount) for liquidity providers.  This will keep the spread low and the available volume high.
• Live editing of posted orders.  Let's say I place a bid at $4.90; then someone instantly posts $4.91 above me.  I'm willing to take $4.92, but to do so (particularly on Intersango) I have to cancel my pending order, create a new one for the same volume, but for a different amount.  It would be more convenient if I could just push edit and have all that process automated.  (this would also be helped by the BUY BTC/SELL USD distinction I mentioned above).

I'm sure I had more, but can't recall off the top of my head.

I suspect that this one is easy to say, but hard to do.  If I were designing the system, I think I'd be basing it all off a log system.  Orderbook adjustments and sales all appended to the same DB table and piped out to the same stream.  The engine's job would simply be to execute lines of the log and push that fact out to the live stream.  I'd also be looking at ways to reverse proxy out the live stream, making it semi-peer-to-peer.  That would hopefully prevent the dreadful performance that Mt.Gox suffers at exactly the moments that things get interesting.


Yeah, I didn't expect that it would happen.  Although on the optimistic side -- Bitcoinica and Mt.Gox are pretty easily transferred between.  I guess it's simply a matter of maintaining an account at Mt.Gox (and any other exchange with a deposit code system).


I may not have used the right word.  I'll give an example.  Consider this orderbook:


Assume I have 1000 BTC.  Let's say I place my order as a BTC SELL of 1000 BTC at a limit of $4.9.  When executed with the above orderbook, that will turn into:

• 750 @ $5 = $3750 (leaving 250BTC outstanding)
• 250 @ $4.95 = $1237.5

I end up with $4987.50; which is more than the (4.9*1000) $4900 my limit order implies (obviously I'm happy, there were people bidding higher than I was asking, so placing my order later than them gets me a bonus of $87.50).  The difference between these two is what I meant when I said "slippage".

Now start again.  This time I place the order as a USD BUY of $4900 at a limit of $4.9.  Given the same orderbook as the previous example, this turns into:

• 750 @ $5 = $3750 (leaving $1150 outstanding)
• 232.3232 @ $4.95 = $1150

I end up with $4900 and 17.6767BTC.

The difference should be obvious.  In one case my "bonus" is in USD, in the other it's in BTC.  At present no exchange agrees with me that this is relevant (particularly the Intersango guys who say there is no difference between BUY USD and SELL BTC).

Here's my use-case: I am a merchant.  I want to sell my widget for $100.  I quote a price in BTC, say 20.40 BTC.  My order system instantly hedges against price changes by issuing a "BUY 100 USD" to an exchange.  $100 is all I want, the rest of my hedge fund can be left as BTC.  Eventually I will receive the BTC from the customer (6 confirmations); but I won't care if the exchange price has changed during that hour.

Obviously the opposite case is possible as well when you are the customer.  SELL USD and BUY BTC are different things.



My pleasure; I am glad to have been of use.  Hopefully the above clarifies some of my suggestions.


Depends on how you make the analogy.  Gold's primary function is "being gold" -- in that it's density and non-erodability make it hard to fake or tarnish.  Bitcoin's primary function is "being two special numbers that make a public and private key".

Gold's secondary function is "looking pretty".  Bitcoin's secondary function is "being compatible with the Bitcoin network".

If you believe in "inherent value", then it is that secondary function that gives bitcoins an inherent value.  It's unusual in that it's value is provided by the computers and software that run the network rather than the magic numbers themselves -- it is definitely value though.