i'm currently studying Zhoutong's model. my conclusion so far is that he's very over exposed. his accts are setup within mtgox AND he's complained about intermittent interruptions in his access to that same acct just like the rest of us. he's also admitted that he's not 100% hedged.
in a massive short squeeze, especially one where he can't get access to liquidate clients accts or hedge quickly enough, Bitcoinica will suffer massive losses. his problem is that mtgox's trading platform is outdated and not fast enough for a large ramp in the price.
now is not the time to be shorting. the bursts in price will come intermittently and unpredictably.
Hi,
I've experienced the 9/11, and several many spikes and crashes in prices. Bitcoinica can completely handle 50% change in price within 1-3 seconds. Mt. Gox's system is so slow that our servers can definitely react before the prices move to an unreasonable level.
It's true that some trades are not hedged 100%, because we have a sophisticated internal matching system. That's the only channel for our profits. The spreads are simply the sum of Mt. Gox fees and market spread. Only when we match orders internally, we can make profits from the trades.
We maintain absolutely zero net position as a whole. Our prices reflect instantaneous liquidity, volatility and they always account the risk of Bitcoinica.
Also, for forced liquidations, we don't have to hedge to recover the losses. For example, someone places a limit sell order at $4, and we expect a short seller to be liquidated at $4.1 to cover. (Assuming same quantities.) Then we can safely assume that both orders will be executed during a major short squeeze. Even if our system can't react at all, Bitcoinica will still make $0.1 profit. Our high volume and large amount of limit and stop orders have ensured this to happen most of the time if hedging is not possible at all.
In reality, we apply all the algorithms together. Our interface is extremely simple and intuitive, but the back-end is not that simplistic. Our servers process more than 50,000 database requests per minute, just to be able to calculate all the metrics in real time.