Typical daily volumes (kBTC) on major exchanges, for the last week or so, from Bitcinwisdom's charts:


Huobi       105
OKCoin       35
BTC-China    11

BTC-e        21
Bitstamp     17
MtGox        16
BitFinEx     16

TOTAL CHINA     151
TOTAL NON-CHINA  70

Do we mean the same thing?

From the 1-day charts at Bitcoinwisdom, bitcoin's October mini-rally first reached ~200 USD briefly on 2013-10-22, then stagnated at or below that level until 2013-11-02, and definitely rose above that level again on 2013-11-03, as the beginning of the big rally.

I could not see clearly which exchange started the rally; on the 1-day charts they all seem to have started together.

It seems  however that the three Chinese exchanges had about 30% of the total volume on 2013-10-22 (the "first time"), and about 50% on 2013-11-03 (the "second time"). 

Volume at the main exchanges on the three days up to the all-time high, namely 27, 28 and 29/nov, when the price was between 950 and 1130 (on Bitstamp; 10% more on MtGox), in kBTC:

BTC-China  80+80+50 =  210
OKCoin     62+55+30 =  147
Huobi      45+37+20 =  102
MtGox      38+37+15 =   90
BTC-e      39+27+13 =   79
Bitstamp   37+22+13 =   72
BitFinEx   9 + 9+ 5 =   23

Other exchanges had smaller amounts.  In total, about 750,000 bitcoins were traded on those exchanges in those three days, when the price was at or above 960 USD and it seemed that it would keep increasing well beyond 1200 USD. 

Presumably some of those coins were traded multiple times in that period, but it seems safe to assume that there must be a few hundred thousand bitcoins out there which were bought at those prices by their current owners.

On that occasion the volume at the Chinese exchanges increased and easily ate up that barrier.  Before that, the price stagnated for 10-12 days  just below 200 USD.  Would it have broken that barrier without the Chinese?  Who knows...

It seems that tens of thousands of coins were bought by speculators at 1000 USD or more, just before the all-time high, on the assumption that the price would keep rising. Some of those speculators must have grown disenchanted and will dump those coins as soon as the price returns to that level.

Even for on-book transactions it would be hard to get the exact price.

In all cases one would have to assume the price from some weighted mean of the prices at the main exchanges, at approximately that time (with proper currency/penalty conversion factors, e.g. 1.00 for Bitstamp, 0.90 for MtGox, etc.)

That table will be very crude of course. A statistic per wallet rather than per coin may be more interesting.  That is, for each wallet we keep an "estimated USD balance", and for each transaction of N bitcoins into (or out of) it, we subtract (or add) N*P dollars, where P is the mean BTC/USD rate at the time.  Then we can estimate the average price that was paid for the coins remaining in that wallet.

Doing that for all wallets may be prohibitive, but it would suffice to do it for a large enough sample.

Of course it would not say much about the position of persons, since a person may have many wallets; but even so it would be an interesting statistic.

One who buys only 900 USD worth of merchandise with 1 BTC obviously does not believe that 1 BTC will be worth 10,000 USD in 2015... 

By analyzing the transactions in the blockchain and matching them with a detailed table of BTC price over time, is it possible to figure out the amount of BTC that was last bought at a given price?

The goal would be a table saying, for instance

     10,000 BTC were last bought at 1200-1250 USD
       7,000 BTC were last bought at 1150-1250 USD,
           ... 
 5,000,000 BTC were last bought at at 0-50 USD.

Such a table should be useful to speculators, no?

Huobi traders, humans or robots, seem to have gone to bed: hourly volume is down by 90% or more.  As happens almost every day at this time (small morning hours in local time).

I agree with your criticism of exchanges.

However, even if cryptocurrencies succeed, traditional currencies ("fiat") will be around for many decades. Governments will pay salaries, pensions, contracts, treasury bonds with fiat, and will collect taxes in fiat. In some countries (as in China now) merchants must quote prices exclusively in fiat.

While fiat exists, exchanges will be needed so that people who are not financial tycoons can convert fiat money into crypto money and vice-versa.

At first, the BTC/USD rate will be set by speculation on the exchanges, and that rate will determine BTC's purchasing power for goods and services.

If and when commerce shifts substantially to drect cryptocoin payments, things may reverse: the market of real goods and services will eventually define the purchasing power of cryptocoins, and that will determine the fiat/crypto rate at the exchanges.

 

I see.  But if the bitcoins actually change hands, then the trades (robot or not) are not fake, and and the price they define is the real price.  Non-zero fees obviously drive speculators away from other exchanges.

If Huobi's daily BTC volume is now really 4x that of Bitstamp, MtGox and BTC-e (as Bitcoinwisdom seems to show) , then they should tend to control the market, with the others following through arbitrage.

Bitcoinwisdom has daily data for them since Sep/2013. (It has several years of data for other exchanges.)

What makes you think so?

If it is fake, it is very well done. It even looks more real than the other exchanges...

Huobi's volume usually falls by 90% during the early morning hours local time (which is UTC plus 9 hours, I believe). However that dip was missing on Jan/07 UTC (Jan/08 local), during the crash from ~5600 to ~4700; and just now, on Jan/10 UTC (Jan/11 local), during the ongoing rally.

Perhaps the rally at Huobi is being driven by non-Chinese customers?

According to the 1h charts at http://bitcoinwisdom.com/markets/btce/btcusd, the current mini-rally apparently started at

  15:00 on Huobi
  15:00 on BTC-e
  18:00 on MtGox
  20:00 on Bitstamp

all on Jan/10 (UTC).

These bogus arguments are still circulating, although the "official bitcoin sources" now seem to avoid them.

Governments can outlaw or regulate Bitcoin as they wish.  Their law enforcement agencies generally have sufficient power and tools to ensure that most citizens will comply with the laws.  See China and India for example.


I cannot recommend to anyone that they invest money in something which, in my opinion, is unlikely to ever be more than a form of gambling.

As for impact, last year we saw in Brazil the collapse of TelexFree, a rather obvious Ponzi schema, and OGX, an oil company which allegedly misled investors about productivity of its wells.  Each involved the transfer of several billion dollars from the pockets of smart people to those of even smarter people. (The TelexFree operators alone pocketed more than half a billion USD; still unclear how much more.)

I do not see the connection between the two.

Right now in http://bitcoinwisdom.com/markets/huobi/btccny I see about 10 transactions per minute, which would be about 15,000 per day and thus about 3 BTC per transaction on average. 

Are they showing the entire order book, or only the N highest bids and lowest asks?

Is is possible that the exchanges are posting fake orders, that will never get closed with real transactions?

I am a dinosaur academic type.

I understand enough of the protocol to believe that it is technically sound.  My skepticism is about the economic aspects.

It doesn't bother me that people are speculating (basically, gambling) on a market with very uncertain future -- as long as everybody understands the odds and risks.

It does bother me that Bitcoin is being "sold" by some with rather misleading claims, such as "the supply is limited" (true of Bitcoin, but not of cryptocurrencies in general), "transfers are anonymous", "cannot be stolen or taxed",
 "there are no fees", "banks and governments cannot touch it", etc. 

And it bothers me a lot when speculators turn to Latin America as a promising source of Rich Suckers -- who, they hope, will buy those overpriced Bitcoins that no longer find buyers in China, India, Europe and North America.

Not if I can help it.

Those are part of (a) or (d) depending on how much faith they have.  The (a)s would have the hassles and losses of converting USD to BTC, a limited choice of merchants, and will have nothing to show for it except the receipt. (A Lamborghini bought through that route looks exactly the same as one bought with USD.)

Those are included (c). (Perhaps I should have said "Miner or Speculator".) 

By the way, spending 1 Bitcoin that is supposed to be worth 10,000 USD next year to buy 900 USD worth of merchandise today is almost a confession that one has lost the Faith.

Please read the rest. Obviously average Joe will NOT care about it. Why would someone convert USD to BTC if they can buy with USD directly?