I just improved the Wikipedia article about MACD.  Then I read it to learn what MACD was. 

Er, um, sorry, but I am not impressed.  The interpretation of those crossing signs sounds a lot like haruspicy. 

Chinese Slumber Method prediction for Thursday April 24

Prediction valid for: Thursday 2014-04-24, 19:00--19:59 UTC (not before, not after)
Huobi's predicted price: 3059 CNY
Bitstamp's predicted price: 487 USD





[ Plot legend ]

Today's data point was rather good, (S = 0.0039 W = 0.742), and not too much below the previous trend line (by 54 CNY).  It seems still valid to use a straight-line trend, fitted by weighted least squares to the last five points, Apr/19--23.  Namely, A + B*(d-d0), where d-d0 is the number of days since Apr/19, A = 3151.11 and B = -18.42.

The Bitstamp prediction, as usual, is the Huobi prediction divided by the currency conversion factor R, which was assumed to be 6.28 CNY/USD. It was 6.28, 6.34, 6.27, 6.25, 6.26 at the last four Slumber Times.
 
Checking the previous prediction

Prediction was posted on: Tuesday 2014-04-22, 22:35 UTC
Prediction was valid for: Wednesday 2014-04-23, 19:00--19:59 UTC (~20 hours later)

Huobi's predicted price: 3112 CNY
Huobi's actual price (L+H)/2: 3058 CNY
Error: 54 CNY (~9 USD)

Bitstamp's predicted price: 494 USD
Bitstamp's actual price (L+H)/2: 487 USD
Error: 7 USD
 
NOTE: "Why should someone who knows something conceal it?"
  -- Sumerian "proverb", from an Assyrian school exercise collection

For deposits, not really; although if the government did not like the bank-transfer channels, they may  not like also any workaround that the exchanges may use to get the same effect (such as Huobi's recharge cards).  But there are no news at present.

For withdrawals, the Caixin article that first leaked the PBoC circular (discounting the earlier leak, denied on the same day by the PBoC) gave April 15 at the deadline to close all ties, after a period when only deposits would be blocked.   Obviously the implementation of the circular has been happening more slowly than claimed in that article; but, apart from that, the article has been confirmed so far.  Thus there is still room for more bad news.


The news I saw were about ONE Australian bank (a government-owned one perhaps?) that decided on its own to close an exchange's account as of May 1.  Are there other similar decisions?
 

I don't follow...

The demand would go up very little if people were buying BTC for the purpose of spending them right away.  Each transaction of this type would decrease the circulating bitcoin supply (CBS) for a short time but that would be reveresed when the merchant or payment processor converts the BTC back to fiat. Therefore, the persistent decrease of the CBS from these uses would be proportional to the (total daily BTC volume V1 in these transactions) x (time T that each of those BTC stays off the system).  As long as transaction volume remains constant, this term will not change the CBS.

However, some of the bitcoin purchases are being made by people who were hoarding BTC and started spending some of them.  Assuming the coins are converted to fiat by selling on market, each of those transactions increases the CBS permanently by the amount transacted.  So these transactions keep increasing the CBS as time passes.  

I would guess that most of BitPay's 100 M$/year volume in 2013 is due to the second type of transaction, because buying bitcoins just for e-payment does not seem to be sufficiently attractive to people who do not own bitcoins already.  If we assume a 50-50 split, and off-market time T of 3 days then perhaps 0.5 M$ worth of bitcoins were taken out of the CBS temporarily by type 1 transactions, and 50 M$ worth of coins were permanently added to the CBS by type 2 transactions last year through BitPay. 

EDIT: clarified "last year"

EDIT2: Clarified "through BitPay". The same analysis should apply for commercial payments that do not use BitPay. Is there an estimate of this volume?

I don't know much, I can only guess.

Deposits via the exchange's bank account appear to have been closed in all exchanges.  Huobi's recharge cards were available for a long time before the latest tightening and probably have limits on amount per day and such. OKCoin's broker/recharge-code system seems to be cumbersome, like BTC-China's voucher system that apparently was never popular.  So, obviously, moving money into the exchanges has become more difficult now.

On the other hand, withdrawing money via bank transfers seems to be unaffected, so moving money out is as easy as before.

My explanation for the general decline since February is that the Chinese speculators are gradually cashing out and leaving, and no new investors are coming in.  (Why would anyone be tempted to invest in an item that he cannot use for anything but speculation, and whose price has been generally falling for almost 5 months?)

Daily volume V at Huobi seems rather low. It is already 08:00 pm in China and V is still less than 18 kBTC.  At this rate, by the time they go to bed it will be less than 30 kBTC.

The last times it was lower than that were

  Sun Mar/23 (18 kBTC) ,when they were DDOSed and stayed offline for 3-4 hours in the mid-afternoon, IIRC;

  Sun Mar/16 (27 kBTC), don't recall is anything special happened;

  The New Year Week Jan/29-Feb/06 (< 18 kBTC), when banks were closed.

Then, presumably,'

Bobby Lee = Chinese gubmint
Those alexeys of BTC-e = Russian gubmint
Mircea Popescu = Romanian gubmint
The guy of Bitstamp = Slovenian gubmint
Antonopoulos = Greek gubmint
Tuur Demeester = Dutch gubmint

And so on.  Geez, the whole United Nations are out to kill Bitcoin! ...

The UK government could stamp out bitcoin much more easily by declaring it private money hence illegal. (I am assuming; every government which is minimally awake should rigorously forbid privately issued currencies.)

That is why it would be interesting to know Danny's history in Britain before going to Cyprus.  Hard to believe that no one in Cyprus bothered to ask, and no reporter in England cared to find out...

Thanks!

Is there an explanation of that plot somewhere? (How can the number decrease? Because empty addresses are excluded?)
Thanks...

Thanks, will check.

Chinese Slumber Method prediction for Wednesday April 23

Prediction valid for: Wednesday 2014-04-23, 19:00--19:59 UTC (not before, not after)
Huobi's predicted price: 3112 CNY
Bitstamp's predicted price: 494 USD





[ Plot legend ]

Today's data point was very good, (S = 0.0016 W = 0.949) and only 29 CNY above the previous trend line.  Therefore it seems appropriate to use again a straight-line trend, fitted by weighted least squares to the last four points.  Namely, A + B*(d-d0), where d-d0 is the number of days since Apr/19, A = 3128.64 and B = -4.27.

The Bitstamp prediction, as usual, is the Huobi prediction divided by the currency conversion factor R, which was assumed to be 6.30 CNY/USD. It was 6.34, 6.27, 6.25, 6.26 at the last four Slumber Times.
 
Checking the previous prediction

Prediction was posted on: Monday 2014-04-21, 23:29 UTC
Prediction was valid for: Tuesday 2014-04-22, 19:00--19:59 UTC (~20 hours later)

Not bad, but admittedly it was not a hard guess:

Huobi's predicted price: 3091 CNY
Huobi's actual price (L+H)/2: 3120 CNY
Error: 29 CNY (~5 USD)

Bitstamp's predicted price: 494 USD
Bitstamp's actual price (L+H)/2: 492 USD
Error: 2 USD
 
NOTE: "We encourage you NOT to read the posts of anyone with whom you know you will disagree."
   (Mindspring.net's canned reply to spam complaints, maybe around 1990.)

You seem to be assuming that the "bandits" will act in the smartest way possible, and that the police will assume that they do so.  To my knowledge, neither is true: bandits often make mistakes, and the police depends a lot on them doing so

But we'll have to wait and see, I guess...

Discouraging crime depends on  the police being reasonably efficient at catching criminals after the fact, more than catching them when they attempt to commit the crime.  If murderers, thieves, rapists could only be caught on the act, they would only have to commit their crimes when the police is not around.

I don't  see how one could stop the continuing epidemics of goxitis -- "hackers stole all your bitcoins, sorry" -- and other forms of bitcoin theft except by tracking the stolen coins and identifying the thieves.   Are you saying that this will not be possible, and should not be?

I don't see why those "proofs" are necessary.  Bitcoin believers will totally trust anyone who claims to love bitcoins and any outfit that deals with bitcoins. 

They even believe that those "proofs" prove anything...

Indeed they are the same thing, the problem is that it seems much easier to tie an innocent person to a "dirty" bitcoin address than to a "dirty" bank account.

EDIT: Note that, in the original scenario, address A1 is created and controlled by X, not by Y.  So it is  not the same thing as depositing dirty money into someone else's legitimate bank account (or dirty coins into someone else's legitimate address).  The crucial difference is that X can take out the bulk of the dirty money from that account.

I wonder if the anonimity of bitcoin addresses could create risks of misidentification of address owners by the police, even for people who have never used bitcoins.

Suppose that a "bandit" X generates a new address A1, deposits a pile of "dirty" bitcoins in it.

X then makes some small purchase from a retailer Z, pretending to be some other person Y who has nothing to do with the matter.  He provides Y's physical address for delivery (if the purchase is a physical item) and pays in bitcoins out of address A1 to the retailer's address A2. Then X moves the rest of the dirty coins from A1 to a tumbler, or whatever.

While following the trail of the "dirty" bitcoins, the police gets to address A1, notices the small transaction to address A2, that they recognize as belonging to retailer Z. From the retailer's record of the purchase, they identify Y and conclude that he must be the owner of address A1 and hence of the "dirty" coins.

For bandit X, the purpose of this this ruse would be to delay the police's pursuit, or even permanently derail it.  He could choose the innocent decoy Y so that the suspicion would seem plausible -- eg. someone not very computer-savy and with criminal background, large debts, drug addiction, etc. 

Many variants of this trick seem possible.  Bandit X could do the same with dozens of decoys, to get the police bogged down into checking dozens of false leads. 

Is this risk significant?  If so, how could it be reduced?

Criminals already use this trick with bank accounts. However, the risk for bandit X is very high in the bank version, because he must steal Y's identity and impersonate him when creating the bank account A1.  On the other hand, the police usually can figure out very quickly that Y did not create that account nor issue the payment and transfer orders.  (Here in Brazil, the innocent victim Y is called "an orange" in fraud jargon, and he is usually some poor semi-literate guy who does not even have a bank account and lives far from the branch where the account is opened.)

With bitcoin addresses, however, the risk for X is very small, since he does not have to get Y's documents nor interact with anyone else to pull the trick, other than issue the bitcoin transaction requests and the bogus purchase order in Y's name; while Y can be any person in the world, and he has no way of proving to the police that he is not the owner of address A1 and did not make the fatal purchase.   His only hope is that the police can trace the path of the purchase order and/or the bitcoin transaction requests on the internet, and determine that they could not have been issued from any computer that Y could have accessed.

Ancient Greeks believed that when the Gods wished to punish a mortal, they would grant his most ardent wish. 

Maybe... It depends on what is the aim of the December decree.  Methinks that PBoC is satisfied with the current situation (bitcoin not used for commerce or by financial institutions), but law enforcement and national security agencies may still see risks in it.
 

Huobi's recharge cards are called "火币点卡充值" in their Chinese homepage (火币="Huobi" 点卡充值="TOPUP" via Google Translate) and "CCYOU" in their English homepage.

I am largely guessing, but they seem to be like debit cards issued by some card company, that can be recharged by sending money to the card company and can be spent only(?) at Huobi.

The OKCoin "recharge code" system seems to depend on one or more OKCoin users with lots of CNY in their OKCoin accounts ("brokers"). As far as I understand, a normal client pays a broker somehow, without involving OKCoin or their bank account, and the broker gives the client a code that he uses to transfer CNY from the broker's OKCoin account to his own.  (In their forum, one client complained that the only broker has a 10,000 CNY minimum so how could one do small occasional trading?)

This system may be usable also by clients outside China who, say, pay the broker in dollars on an US account.  I note that a substantial part of OKCoin's volume goes on 24/7, unlike Huobi's that stops at night.

PS. However, Huobi's "recharge cards" apparently can be recharged by bank transfer, so in practice the bank transfer channel is still open but indirectly.