In normal conditions there is a fairly stable ratio between prices at Huobi and prices at other functioning exchanges, such as Bitstamp; especially  at the "Chinese Slumber Times" when nobody is stirring at Huobi, not even a mouse.

Presumably the ratio is maintained by arbitrage traders.  It is a bit higher than the official exchange rate (1 USD = 6.09 CNY) and it seems to change sometimes.

For the last two weeks the effective ratio R = (Huobi price)/(Bitstamp price) it has been quite constant at 6.12 CNY/USD, except for the three days after the New Year holidays (feb 07--09) when it changed to 6.40.

(When the Chinese banks reopened, the price at Huobi started to fall.  Bitstamp's price did the same, but at a faster rate.  Maybe it was over-reacting to the Chinese moves, or trying to antecipate them.)

As long as this ratio R is maintained, predicitions of Huobi prices can be translated to Bitstamp prices.

In the figure below, the red/green price bars are actual Bitstamp prices, but the yellow dots are the Huobi prices at 19:00 UTC every day, divided by the appropriate R factor (6.40 for feb/07--09, 6.12 for the other days).  Note that they fall quite accurately on top of the actual price plot.



(In this plot I used 19:00 UTC (4:00 03:00 am China) as the nominal "slumber hour", whereas in the previous plot I had used 20:00 (5:00 04:00 am). Sorry for any confusion.)

The blue rectangles are the three predictions I made previously, based on the Chinese Slumber Method and the expected R factor.:

 * The first prediction was right on target. It used the factor  R = 6.40, estimated from the prices at Huobi and Bitstamp for the previous two days.
 
 * The second prediction too used R = 6.40, as it had been for the last three Slumber Points.  It  was foiled by the Karpeles Catastrophe, but that error was partially compensatd by the R factor unexpectedly returning to ~6.12 after the Catastrophe.

 * The third prediction would have been spot on (even more accurate than the one for Huobi)  if I had followed my Chinese Slumber Method blindly, namely the Huobi prediction divided by R.  But the result looked so unlikely that I decided to adjust it based on where the Bitstamp prices seemed to be going.  Stupid of me.

Let's see if the method works again for tomorrow... Prediction coming next.

Those "Chinese Slumber" trends that I believe I see in the Huobi hourly price charts are rather unexpected, since market prices usually behave as if they had no memory of the past, only of the present (that is the principle of the "geometric Brownian" model and its fancier variants.)  So, after wandering randomly for 24 hours, the market should not be influenced for where it was 24 hours earlier.

But I believe that those trends are real, and are a consequence of trader habits and of the relatively closed nature of the Chinese market.   

During the day, traders will move their holdings back and forth from Yuans to Bitcoins, often many times per day, and will move Yuans between the exchange and their own bank accounts.  But presumably every trader has a preferred "base position", a ratio of Yuan to bitcoin that he would rather leave in his account while he is not watching, and (especially) the total amount of money he is willing to leave in the exchanget.

The "base position" of different traders may be very different, but those differences and details don't matter. The important point is that before going to bed, each investor presumably tries to return to his "base position" if he can do so without much loss.  I believe that the net effect of these tendencies, averaging out all traders, is that the total amount of CNY (tCNY) in the traders' accounts returns to about the same value every night as everyone goes to bed.

On the other hand, the total amount of bitcoins tBTC in the exchange should vary slowly, too. The typical Chinese trader should have little motivation and/or means to trade at exchanges outside China, and the amount moved by arbitrage traders (even between Huobi and OKCoin) should be a small fraction of tBTC.

The next step in the theory, that is assumed proved by a suitable amount of hand-waving, is that there is a strong connection between the ratio tCNY/tBTC and the market price of 1 BTC in CNY at that exchange.

With those hypotheses, it follows that the current BTC/CNY prices, observed every 24 hours at the "slumber times", should vary much less than the hour-to-our prices, or prices taken every 24 hours during daytime.

In particular, during the New Year holidays, Huobi was effectively a closed market,  with no inflow or outflow of Yuan.  Thus tCNY was constant, and the price (especially at the slumber times) remained practically constant.

Once the banks opened on Feb/07, there was both temporary flow of money into and out of the exchange during the day, leading to highly variable prices;  but also a smaller and relatively steady net outflow of money, as speculators became disappointed with their returns, cashed out, and left the market.  The prices at the slumber times decayed because of this net outflow, which reduced the tCNY total.  Between Feb 07 and Feb/10 he price (and presumably the tCNY) seemed to decay as a shited exponential, tending towards some limit in the 4000 CNY range, implying that the net outflow of CNY was decreasing exponentially towarsd zero.

However, that trend was interrupted by the Karpeles Catastrophe on Feb/10.  I epect that the net outflow of CNY from Huobi will be restarted, but again decrease exponentially with time;  so perhaps the slumber points will again follow a shifted exponential trend.

My last prediciton for Huobi by the "Chinese Slumber" method was a bit too optmistic.

Prediction made at Feb/12, 02:40 UTC
Prediction valid for Feb/12, 20:00 to 20:59 UTC
Huobi's predicted price: 4045--4105 Yuan
Huobi's actual price: 4000-4028

See the plot below.  My last three predictions are marked by blue rectangles.


The "Chinese Slumber" method looks at the price only at those times when most Chinese traders are asleep, which is around 19:00--20:00 UTC (04:00--05:00 03:00--04:00 local China time).  Those prices and times are marked by yellow dots on the plot.

Over the last two weeks, if not more, those dots seemed to follow fairly regular trends, while between the dots (when the Chinese were trading) the price often went wild.  The "Chinese slumber" method simply extrapolates the current trend to guess the next yellow dot.

The method worked beautifully the first time, using a shifted-exponential trend that fitted the previous 3-4 dots.

The second prediction was totally elmerfudded  by the Karpeles Catastrophe.

For the third prediction, I rashly extrapolated a new trend with only two dots.  Furthermore I could not get a clear reading of the price at the second dot because the Chinese (or perhaps arbitrage robots?) kept trading all through the night at Huobi.   It seems that my reading was too high.

I will keep trying...

(I have a theory for a possible cause of those Chinese Slumber trends, will post it later.)

EDIT: image redone to use 19:00 UTC (4:00 03:00 am China) instead of 20:00 (5:00 04:00 am)  as the nominal Slumber Time for the yellow dots.  Also corrected the yellow dot of Feb/11 to show the price I had used in the last prediction.

EDIT: fixed local times (UTC+08:00 not UTC+09:00)

And bitcoincharts cryptocoincharts STILL pretends Huobi does not exist, even though is twice the size of OKCoin and three times the size of Bitstamp by daily volume.

They should not publish charts that are so grossly wrong.   Bitconwisdom gets Huobi's data with no problem, why can't Bitcoincharts cryptocoincharts?

I have been told that some bitcoiners hate Huobi because they blame them for the December crash.  Methinks they are to "blame" for the price not having fallen back to 17 US$ then...

EDIT: site's name.  Bitcoincharts still ignores Huobi AND OKCoin.

Looking at the historic price charts for the last few years, I would think that China is responsible for 95% of the current price.

Specifically, it seems that the Feb-Apr 2013 price increase, from ~17 US$ to ~100 US$, was due to the opening of BTC-China in Hong Kong; and the Oct-Nov rise from ~100 US$ to ~1100 US$ followed the opening of Huobi in Beijing, aided by OKCoin which had opened a few months earlier.

This explanation would make economic sense, since the opening of a new large market for a product with fixed supply should make the price rise.

Further evidence of China's predominance is that the price fell to almost a half when the Chinese government started restricting the use of bitcoins, and partially recovered when Huobi and OKCoin found that they could still do currency deposits and withdrawals through a Chinese bank.  In contrast, similar restrictions by other large countries without significant exchanges (such as Russia, India, and Indonesia) do not seem to have had a significant impact on price.

Finally it is evident that the prices in all functioning exchanges, inside and outside China, are closely tied together, so that any significant change in one is echoed in all the others within a few minutes.  Presumably this is due to arbitrage trading.  From what I understand, arbitrage between two markets should have the greater influence on the market which has the lower liquidity, which is usually the one with smaller volume.  That is, the larger markets tend to drag the smaller ones, as one would expect.

Huobi's CEO recently stated that they have about ten thousand actively trading clients.  One cannot trust this sort of data, except as an upper bound: the real number may be much less, but not much more.   It would be nice to have similar estimates for the other exchanges.  (We have their trade volumes, but since Huobi has no transaction fees, and the others have, one cannot simply use the same ratio.)

It would be nice also to have the total coins and cash in the client accounts of each exchange, as well as in the exchange's reserves.  These numbers should be basic transparency requirements in an eventual regulation of the exchanges, whether by government or by the community.

Can't resist, sorry:

"Our software did not have the MtGOX bug, we have already removed it, and we are suspending bitcoin withdrawals while we work on it."

 

An event that would change the way we view the world?

Are they referring to MtGOX's press release?

 

My predictions for Wednesday February 12, at precisely 20:00 UTC:

Huobi: 4075 plus or minus 30 CNY.
Bitstamp: 640 plus or minus 15 USD.
Sagittarius: Chances are you may have to get out of bed in the morning.  Venus in Ursa means a good day for counting wild mushrooms and sharpening pencils. Beware of talking frogs and freshly painted doorknobs. Someone from Taurus may ignore your posts, or turn them into internet memes.

Daily volumes of BTC trade to/from USD and other national currencies (in kBTC):


             !   Sat !   Sun !   Mon !   Tue !   Wed !    Thu !    Fri !    Sat !    Sun !    Mon !    Tue !                     
  EXCHANGE   ! 02/01 ! 02/02 ! 02/03 ! 02/04 ! 02/05 !  02/06 !  02/07 !  02/08 !  02/09 !  02/10 !  02/11 ! Currencies considered

  Bitstamp   |  8.30 |  4.07 |  4.93 |  7.59 |  8.48 |  20.65 |  42.79 |  13.50 |  15.06 |  71.86 |  40.06 | USD                 
  BitFinEx   |  4.54 |  3.15 |  3.45 |  3.36 |  5.22 |  21.11 |  66.36 |  11.12 |  12.07 |  41.98 |  39.54 | USD                 
  BTC-e      |  6.30 |  4.74 |  5.75 |  4.51 |  6.55 |   9.72 |  31.84 |  11.67 |  12.51 |  47.04 |  28.64 | USD,EUR,RUR         
  MtGOX      |  5.13 |  2.28 |  3.37 |  5.40 |  8.20 |  21.54 |  58.89 |  26.82 |  23.48 |  50.07 |  18.70 | USD,EUR,GBP,AUD,JPY 
  Kraken     |  0.23 |  0.15 |  0.21 |  0.29 |  0.23 |   0.59 |   1.45 |   0.37 |   0.45 |   1.45 |   1.11 | EUR                 
  Bitcoin.DE |  0.33 |  0.35 |  0.51 |  0.27 |  0.31 |   0.61 |   1.30 |   0.37 |   0.47 |   1.54 |   0.67 | EUR                 
  CaVirtEx   |  0.24 |  0.20 |  0.08 |  0.17 |  0.18 |   0.39 |   1.07 |   0.57 |   0.23 |   0.77 |   0.53 | CAD                 
  CampBX     |  0.23 |  0.16 |  0.10 |  0.12 |  0.16 |   0.21 |   0.82 |   0.27 |   0.24 |   0.37 |   0.11 | USD                 

  SUBTOTAL   | 25.30 | 15.10 | 18.40 | 21.71 | 29.33 |  74.82 | 204.52 |  64.69 |  64.51 | 215.08 | 129.36 |                     

  Huobi      | 28.86 | 26.27 | 24.77 | 18.65 | 19.55 |  39.97 | 131.91 |  77.00 |  64.08 | 134.09 | 120.08 | CNY                 
  OKCoin     | 20.38 | 20.00 | 18.09 | 12.77 | 12.33 |  26.07 |  74.37 |  37.63 |  34.63 |  70.49 |  64.22 | CNY                 
  BTC-China  |  2.31 |  2.21 |  1.70 |  1.33 |  1.54 |   4.09 |  17.66 |   6.86 |   5.75 |  18.23 |  11.17 | CNY                 

  SUBTOTAL   | 51.55 | 48.48 | 44.56 | 32.75 | 33.42 |  70.13 | 223.94 | 121.49 | 104.46 | 222.81 | 195.47 |                     

  TOTAL      | 76.85 | 63.58 | 62.96 | 54.46 | 62.75 | 144.95 | 428.46 | 186.18 | 168.97 | 437.89 | 324.83 |                     



All numbers were collected by hand from the site http://bitcoinwisdom.com. Beware of possible errors.

For each exchange, the numbers include only the trade volume to/from the currencies listed in the rightmost column. Trade between BTC and other cryptocoins, such as LiteCoin, is NOT included.

Coinbase is said to use Bitstamp for currency conversion.

Dates on the header line are UTC. Specifically, "01/15" means "from 01/15 00:00:00 UTC to 01/15 23:59:59 UTC". (Beware that Bitcoinwisdom uses your local time, so the date may appear to be off by 1 day.  For example, if you are 2 hours west of Greenwich, it may show "01/14 22:00" when the UTC time is "01/15 00:00".)

The exchange Crypto-Trade was dropped from Bitcoinwisdom's charts on Feb/10.  Its volume used to be around 10 BTC/day.

Thanks for the vote, but see my signature... 

But since I have been mentioned: methinks it would be better and faster to dissove the Foundation and create a new Bitcon Istitute or something, with a completely new board.  The world won't easily forget who the BF founders were, especially if some of their close friends remain on the board -- even if these people are 100% "clean" themselves.

(Didn't Shrem brag that he would only hire people whom he had got drunk or stoned with?  Does that apply to the Foundation too?)

I see.  Well, that varies a lot actually.  Depends on trade volume (which may vary 100x between daytime and nighttime), whether the price is stable or swinging, etc.  The same holds for other exchanges, by the way.

It is ~9:00 08:00 am now in China, and the price has been very unusual since yesterday.  

I still don't understand what problem you see.

You are aware that the first two columns (detailed order book) are independent of the other two (summary of the order book), right?  It is not one table with four columns per row, but two tables with two columns, and the rows of one are not related to the rows of the other.

On the first two columns, going up from the spread to the 4100 CNY line there are several sell offers that seem  to add up to 67.  Below, on the bid side, the entries less than 4050 add to less than 1 (hence the "0" on the summary) and the next ones down to 4000 seem to add to 9.  Those coins have not been traded yet; they are waiting for matching offers.   The transaction log (trades already executed) is a separate table, below the part you have copied.

So what is wrong?

I don't know if I understand the question. Is it about the current price (large green number)  being above the lowest ask?

This may look wrong, but the current price is the price of the last transaction that was executed.  Since that time, a new sell offer was added to the order book, below that current price.

As I wrote before, I can't understand why the bitcoin community still puts its trust on a Foundation that had Shrem and Karpeles among the 3-4 founding and most important members.

I may be too cynical, but the MtGOX malleability bug reminded me of a scene in the movie /Fargo/ where an" car salesman who got himself in an ugly financial mess tries to buy time by sending blurred faxes of sales records to an impatient auditor.

By the way, someone here mentioned the Bitcoinica affair.  I tried to read about it, but the least news I found on the net, from mid-2013 perhaps, said that the Bitcoinica liquidators were unsucessfully trying to get from Mark Karpeles the number of BTC remaining in Bitcoinica's account at MtGox.  How did that affair end?  (From the original post I understood that those coins may be still at MtGOX.)

Wait, I had understood that there were two massive hacker attacks going onright now: a huge DDOS attack against some unspecified company's site and a lareg-scale spamming of the Bitcoin network, specifically, with "malleated" copies of transaction messages.

The potential vulnerability of exchanges to "malleated" transactions was a separate concern that (as far as we know) was exploited only at MtGOX, over many months, but prompted many other exchnges to udate their software over the last 24 hours.

Isn't this correct?

Huobi now below 4000 CNY, first time since Dec/21.

That's one half of its all-time high of 8000 CNY on Nov/19 (although that was a fluke and should not be counted, the legitimate ATH was 7550 CNY on Nov/29.)

It varies day to day.

Yesterday: Huobi ~134 kBTC,  Bitstamp ~72, OKCoin ~71, MtGOX ~50, BTC-e ~47, Bitfinex ~42, BTC-China ~18.
 
Bitfinex was the largest outside China on Friday Feb/07 (~67 kBTC).

世界中国的中心 !

People usually place sell offers because they just want to sell. 

The higher you set the offer to sell, the longer you will have to wait before you can use your money or play another turn at the game.

The BTC price on MtGOX now (580 USD) is already less than half of its all-time high on Nov/28 (1242 USD).

On other exchanges it is still somewhat above 50%.