Has anyone thought about the potential of a daily crossing network ? -- these are very popular in conventional finance for increasing liquidity and allowing large trades to go through without impacting the market.
The basic idea is that some people are trying to trade large amounts of BTC for USD but don't want to a)show their hand or possibly b)get a partial execution. It's quite possible that with the price over $4 there are people out there who would like to sell say 10,000BTC if they could get $4 , but don't want to clear the MtGox order books out, kill the price and get a partial fill.
Equally, there are possibly some folk who would like to buy substantially at say $5.
In both cases, these people want to trade size, and don't want to work it through the public books, ideally, if they can find each other then a trade can be printed at a beneficial price to both of them. If they cannot find each other, then no information is disseminated -- the whole point is that these buyers and sellers don't wish to show their hands unless they can trade.
The benefit to the market is that there is more liquidity, and there is less ripping through the order books. Less volatility, and Bitcoin becomes a better, more stable, acceptable market.
A cross could take place once a day at say 1400 GMT, to try and coincide with most eyeballs being open. Upto 1345, people would submit their orders (suggest MtGox / Tradehill / Bitcoinica as they need to hold both Fiat and BTC on each order). At 1355 the network would close, and the the optimal clearing price (if any) is computed. This is the single price at which most Bitcoins would be bought/sold given the participants' criteria. Trades would then be printed at this price (minus a tiny commission for the network). The cross would then be over until the following day.
As an example imagine that MtGox is showing 4.5-4.6 with size 25x10
The cross accepts the following orders:
A : P 1000 @ 4
B : P 60,000 @ 6
C : S 200,000 @ 4 (all or nothing)
D : S 5000 @ 4.5
The key-point here is that NO-ONE sees these orders except the crossing computer, C is a huge seller which will spook the market but he has indicated all-or-nothing, and although his price is keen, he will not participate.
The crossing price in this example will be mid-way between B and D i.e. 5.25, with a volume of 5000. This is the only information that is released to the participants, which can be seen as a daily fixing price.
If C had not insisted on such a fill-or-kill order, then the crossing price would instead be 4 where 61000 can trade. D would not participate. C may of course re-enter the following day, only he knows his true size.
Anonymity is the key issue here, this guarantee will attract additional orders, and equally, large speculative orders can be worked safely away from the main books.
A regular and well delivered cross would gradually grow in importance and add liquidity without impact to real time Bitcoin prices.
I like this idea because of the following:...kill the price and get a partial fill.
