BTC-China operated on a voucher system exclusively for some time after the december PBoC decree. Its volume was very small and seemed to be only arbitrage.  They only started to recover after they enabled bank deposit/withdrawal.

I suppose that they are preparing for a closure of the mainland bank accounts?

So the owner of an international account would buy international BTCC vouchers from some outlet that sells them, then send them (digitally I presume) to BTC-China.  Or vice-versa.

Presumably those international vouchers cannot be bought or sold in China.  Otherwise I suspect that the PBoC will not let them do it.

Google search on the thread Wall Observer BTC/USD - Bitcoin price movement tracking & discussion returns pointers to the wrong pages.

A user recently deleted all his posts (hundreds?) on that thread. Presumably messages moved to different pages, and the Google database needs to be updated/rebuilt.

I had an example worked out but I cannot find it now, both Google search and the Forum's search seem to be broken.

Anyway, suppose at the end of the month the arbitrager's Bitstamp account gained 10,000 USD while the Huobi account lost 20,000 yuan (so the two together made a profit).  Then (to simplify the math), he waits for a moment when two exchanges are momentarily in equilibrium, that is, the price at Bitstamp is x and the price at Huobi is 6.21*x, and have good liquidity. Then the arbitrager buys N BTC at Bitstamp (slowly so as not to disturb the equilibrium), sends them to the yuan account, and sells them there (ditto); where N is such that +10000 - N*x = -20000/6.21 + N*x.  Then both accounts will have made the same gain.

This rebalancing can of course be done simultaneously with the arbitrage, even out of equilibrium, so that the two accounts have the same profit at all times.

That possibility seems rather unlikely, but since we are discussing it...

I don't  see how he would notice the error.  Say, his robot would buy some coins at Huobi for less than 6100 CNY and sell them at Bitstamp for more than 1000 USD; that way he would make more proft than he thought, but his robot would miss some opportunities: buying at 6180 would still be profitable but the robot would not think so.  Or, conversely, the robot buys at slightly less than 1000 at Bitstamp and sells for slightly more than 6100 at Huobi; then he is actually losing money, but will only notice if/when he tries to withdraw any excess CNY at Huobi and exchange them for USD at the banks --- which he doesn't have to, since he can rebalance the two accounts by moving bitcoins, at the artificial rate of 6.10 CNY/USD that he himself established.

Note that once the coins are sold, he will not know or care about what happens to them; even if they are bought by another arbitrage robot who sells them back at the other exchange.  (Trading fees will probably prevent a runaway loop there.)

So, on the balance, I think his buggy robot would just miss some opportunities, which a competitor could exploit; but he may never notice that. 

Who knows, perhaps there is only one arbitrage algorithm working between China and the West, with the exchange rate "6.10" hard-coded, and the owner did not notice that it has changed. 

The official exchange rate 

  was 6.06 from Jan/03, 2014 until Feb/14
  climbed gradually to 6.14 through Feb/28
  stayed at 6.12--6.14 until Mar/14
  climbed gradually to the current 6.22 through Mar/20
  has been around 6.20--6.22 since then.

Apparently arbitragers have their own "exchange rate" that has been more or less fixed at ~6.10 since Jan/15.
Until Mar/14 that was not far from the official rate.  It is also possibe that bitcoinwisdom was slow to update
the exchange rate in their scripts. 

This is a plot of the ratio (Huobi price)/(Bitstamp price) daily at ~19:30 UTC, when Huobi volume is usually minimum:


   


The bitcoinwisdom (and other?) charts apparently use the official currency exchange rate 6.21 CNY/USD to compute the equivalent USD price.  However, as the plot shows, the actual ratio maintained by arbitrage is around 6.10 CNY at Huobi = 1 USD at Bitstamp.  It may have to do with Bitstamp trading or withdrawal fees, or other factors.

Humans like round numbers; algorithms don't care for them.   As trading becomes more robotized, those round numbers should lose their "power".

Also, arbitrage between different currencies (euros, yuan, LTC...) will turn round number barriers in one exchange into non-round barriers in another.

Also, I suppose that even a human trader may use non-round prices to get ahead of other traders' round orders.  And he may spread a large order into many small orders, a few dollars apart, to prevent others from doing the same to him.

SecondMarket's Bitcoin Investment Trust (BIT) does not seem to be a large buyer.  According to this thread it has now almost 100,000 bitcoins, including a "seed investment" of about 18,000 BTC.  About 53,000 were bought in 2013; ~4000 in January 2014, ~10,000 in February, ~13,000 in March.

They had promised "liquidity will start no later than March 2014", now changed to "in April 2014, on a limited basis".   Their nominal share value is tied to the price of bitcoin; AFAIK, that is what they will pay to investors who chose and can liquidate.  Thus they could make a good profit they were lucky to sell coins before this crash, and the crash lasts until clients liquidate.

The market obviously is uncertain on the ban being real or not.  There are no confirmations and all media pointedly refer to Caixin as the (only) source. 

However one cannot easily dismiss it as a hoax:

* The reporter is not anonymous, stands by the report, and Caixin seems to be a respectable source.  Moreover the reporter claims to have seen the document, not just heard about it.  And his middle name is not "McGrath"  .

* I have not seen anyone question the plausibility of the ban or its details (e.g. the 15-day withdrawal-only period, and the explicit inclusion of the CEOs' personal bank accounts).  If it is a hoax, it seems to be carefully crafted to sound very "PBoC-like".

* The PBoC had one full day to issue a denial, but did not, in spite of much asking and media attention.  In contrast, when similar rumors circulated on Mar/21, the PBoC denied ir on the same day through its "twitter" (weibo) account.

* When this new report was first published, OKCoin's CEO emphatically dismissed it as a hoax and condemned those who use such methods etc.  I have not seen any more statements like that, it seems that the exchange operators are now unsure, too.

* The rumor is repeated on the website of China's main state TV network, and presumably was bradcast by them.  It seems strange that they would report a leaked PBoC document without checking and getting an OK from the bank.

On the other hand, the Mar/21 rumors and the Caixin article do seem to refer to the same thing, a "strenghtening of regulations" decided by the PBoC around Mar/18.   Perhaps the document shown to Caixin was only a working draft, that in the end was not approved by PBoC.  Or perhaps the ban was still under discussion on Mar/21 (hence the denial), but is now approved (hence the lack of denial).

Or perhaps the reporter was duped by a corrupt official.  Besides the traders who long for cheap coins, I can imagine a couple other ways in which a large temporary price drop would be extremely profitable to certain people.

On the other hand, if the ban is real, perhaps it was leaked on purpose, to soften the expected reaction from the bitcoin trading community.  Namely, if the ban is officially confirmed  on Monday, most traders will be already half-resigned to it.

I found that link too and posted it earlier here:

Check my last two predictions (blue squares at top) 

Another recent Chinese article; seems to say that they contacted the PBoC but got no reply, contacted banks and exchanges but they got nothing yet:
http://tech.sina.com.cn/i/2014-03-29/04419281954.shtml

Chinese Slumber Method prediction for Saturday March 29

Prediction valid for: Saturday 2014-03-29, 19:00--19:59 UTC (not before, not after)
Huobi's predicted price: 2858 CNY.
Bitstamp's predicted price: 470 USD.


Plot legend

On the last two nights (Mar/27 and 28), after the last crash, Huobi saw significant volume at the Slumber Time (S = 0.008 and 0.004, respectively) but the data points are still usable (W = 0.28 and 0.75). Thus it seems reasonable to assume that they define a new straight-line trend after the crash, namely A + B*(d-d0), where d is the day of the month, d0 = 27, A = 3126, B = -134.

The Bitstamp prediction, as usual, is the Huobi prediction divided by the currency conversion factor R, which was assumed to be 6.08 CNY/USD (it was 6.08, 6.03, 6.12, 6.10 at the last four Slumber Times).

NOTE: The dark background seemed more appropriate for the candleflame icon and for these dark and stormy times.

Well, if the last rumor is true, the exchanges will close; and then the secondary businesses will die too, presumably.

Until a few weeks ago Huobi had more volume than OKCoin, now it is a close second place.

here is an interview with OKCoin's CEO http://www.coindesk.com/okcoin-ceo-expansion-china-competitive-edge/

The first article I read from him was glowing, he said he was putting all his savings into it and he loved its supposed anonimity and privacy.  He apparently put all his money in MtGOX, and was quite upset a their failure, which cost him a few tens of thousands of euros.  Then he was disappointed when he figured out that bitcoin is not anonymous at all.  But he is still a fan of bitcoin it seems.

Currently it is not, because the few things that I wish to pay through the internet do not accept bitcoin, but accept credit cards and bank transfers.

I will not buy a life-size rattan elephant or something at Overstock just to make the bitcoins his CEO owns more valuable. 

And if I had invested in bitcoin when I first learned of it (mostly through Rick Falkvinge's rave), I would be very sorry and angry now.

"The Times 03/Jan/2009 Chancellor on brink of second bailout for banks"

I saw it now.  Intriguing, but it is just the front-page headline of the main UK newspaper.

According to the source article, the obvious purpose of that string is to prove that the block was created after that date -- a standard trick often used with photos and videos.   Nothing indicates that the newspaper and date were chosen for the headline.

On the contrary, if the intent was to leave a political message, it should have been more explicit, no?  (Is bitcoin's purpose to make bank bail-outs easier, without having to wait on the decision of government officials?)