A descriptive sum-of-bubbles model for the price of bitcoin
https://bitcointalk.org/index.php?topic=1034430.msg11160620#msg11160620


[ Click on the image for a full-size version ]

Preliminary report - version 2015-05-04. Edits from previous versions:
* 2015-04-22 Fixed caption of Figure 4, added Figure 5.
* 2015-04-23 Added the section "Justification of the model"
* 2015-05-04 Better tuning of parameters (no more rising tails).  Extended price series to 2015-04-20 and added one more "standard" bubble covering the last month.  Added unsmoothed plots (Figures 2 and 8 ).  Added section "Modeling the 2014--2015 price variations" with a 20-bubble model (in blue).


A descriptive sum-of-bubbles model for the price of bitcoin

Figure 1 below shows that the price history of bitcoin can be described fairly accurately by a model consisting of the sum of several idealized bubbles, each consisting of an exponential rise, an optional flat plateau, and an exponential tail.


[ Figure 1. A bubble model for the price of bitcoin, showing the smoothed actual price (grey), the modeled price (green), and the individual bubbles.  The brown line near the bottom is the ratio of the actual price to the model price.  Click on the image for a full-size version. ]


[ Figure 2. The same data and model in Figure 1, without smoothing. Click on the image for a full-size version. ]

Figures 3--6 below show three of those bubbles in linear scale.  Note that, on a log scale plot, such as Figure 1, the rise and tail of a single bubble are plotted as straight lines, but they are distorted into curves when added to other bubbles.


[ Figure 3. The bubble "2011-06" that peaked around June 08, 2011. Click on the image for a full-size version. ]


[ Figure 4. The bubble "2012-08" that peaked around 2012-09-09. Click on the image for a full-size version. ]


[ Figure 5. The bubble "2013-04", with a plateau from about April 5th to May 31st, 2013. Click on the image for a full-size version. ]


[ Figure 6. The bubble "2013-11" that peaked on November 28, 2013, 2013 (the all-time high). Click on the image for a full-size version. ]

The above model has 11 such bubbles.  The first one (red) accounts for the price between 2010-07-15 to 2010-09-30, and is essentially flat at about 0.06 $/BTC in that period.  All other bubbles have a relatively fast ascending rise, with the price increasing between 2% and 10% per day, and a tail that decays more gradually. Four of them have a flat plateau.

In the mathematical model each bubble extends over the entire range of dates considered, but is plotted only while it makes a significant contribution to the model price, specifically at least 5% of it.  Under this criterion, each bubble effectively starts to be relevant and noticeable a few months before the peak, and stop being relevant a few months after the peak.

The brown line in Figures 1 shows that the difference between the model price (green) and the actual smoothed price (grey) is rarely 50% down or 100% up, which is actually quite good considering that the price varied by a factor of about 20'000 = ~2'000'000% (from ~0.06 to ~1200 $/BTC) in the span of that plot.  The largest diferences occur just after the peaks of some bubbles (due to the strong oscillations in the price which are not represented in the model), during the rise of the "2011-06" bubble (which may be two nearly coincident bubbles), and in 2014--2015 (when there were large swings due to the MtGOX collapse, and to rumors and anti-rumors about tightening of the Chinese government decrees).  The 2014--2015 swings can be accurately modeled by adding a few more elements to the model; see below.

Justification of the model

There are infinitely many ways to approximate a function by a linear combination of "basis" functions. One may ask what is special about this bubble basis in particular.

For starters, the bubble basis elements are everywhere positive, and the least-squares fit uses them with positive coefficients.  Other general purpose bases -- such as Gaussian humps, wavelets, B-splines -- would probably require both positive and negative coefficients .

Moreover, the bubble basis above yields a fairly god unweighted least-squares fit, with small relative error, even though the target function varies by a factor of 20'000 over the domain considered.  Usually, least-squares approximations would spread the absolute error evenly over the domain, which would result in errors of tens of dollars even in the early years where the price was below 1 dollar.  (The plot above uses weighted least squares to improve the fit in the early years; however, the unweighted least-squares approximation, not shown, is nearly indistiguishable from that plot, except in the first 3 months Jul--Sep 2010.)

The basis is also very economical (only 11 elements with 45 total parameters) and is directly determined by salient features of the price series.  There are ten salient "bubble-like events", characterized by a period of rapidly increasing price, lasting for a month or more and raising the price by 50% or more, that ends abruptly, creating an evident "kink" in the log-scale plot. (The last two, peaking in early June 2014 and April 2015, are somewhat anomalous and less definite, as discussed below; but they still can be included in this description.)  Just before each kink, the log-scale plot is approximately straight and ascending, which means nearly-exponential growth. In all those ten cases, the kink  is followed by a descending line, initially straight; either immediately, or (in four cases) after a period when the price is nearly constant (once one discounts the influence of nearby bubbles).

After rising from January to late April 2013, and oscillating for about one month, the price spent another month at about 120 $/BTC, then gradually dipped between early June and July, gradually recovered by late August, and remained stable at about 120 $/BTC again, until October 2013, when the next bubble started.  This sequence may have been a single bubble-like event spanning those 10 months, with a temporary remission in the middle of its plateau.  However, it can be modeled fairly well by two flat-top bubble functions, with about the same amplitude.  The start of the sencond bubble's tail is not discernible in the data; it was arbitrarily set to begin around the date of PBoC's first decree (early December 2013) and decay with the same rate as the other Chinese bubbles.

There was also a sharp rise starting in 2014-05-20, that lifted the price fron ~450 $/BTC to ~660 $/BTC in about 10 days.  While the peak of this "mini-bubble" was not as sharp as that of other bubbles, the rise and subsequent slow decay can be modeled fairly well by a bubble function "2014-06" with the same rise and decay rates as the main Chinese bubble "2013-11".  Another similar mini-bubble "2015-03", also with approximately the same rise and decay rates, may have started in February 2015 and peaked on  2015-03-08 (althoug this identification is still quite uncertain.)

The basis merely includes one element for each of those bubble-like events, plus a constant term for the part before the first bubble (Jul--Sep 2010). Note that each pronounced discontinuity in the data calls for at least one basis element with a similar discontinuity.  The typical basis element has four adjustable parameters: its magnitude (the coefficient of the linear combination), the date of the peak, and the rates of the two exponentials. The four bubble elements with a constant section have one additional parameter each, the duration of that section.  Thus, the model has 1 + 4 x 10 + 4 = 45 adjustable parameters in all.

Actually, the rise rates of the last three bubble functions were constrained to be equal while fitting the model, and so were the decay rates of the last four.  These constraints were motivated by the conjecture that those bubbles were connected to the Chinese market (see below), and that their decays were equally affected by the Central Bank restrictions starting in December 2013, right after the all-time high.  Therefore, the number of adjustable parameters was only 40, not 45.

Moreover, the date and rate parameters of each basis element were chosen to match the date of the visible kink(s) of the corresponding event, and the slopes of the log-scale plot just before and just after those kink(s).  It is remarkable that the resulting model also approximates remarkably well the concave curved sections of the price plot between successive bubble events --- even though it has no extra parameters to control the shapes of those curves.

Finally, the basis elements can be interpreted in terms of real-world events and processes (see below); and the composition of the elements by addition (rather than, say, multiplication) is compatible with that interpretation.  Note, in particular, that the curved sections of the log-scale plot arise naturally from the additon of simple exponentials, each of them being straight on that plot.

As noted above, the "bubble event" that peaked in early June 2014 was somewhat anomalous, because the rise phase started quite abruptly on 2014-05-20, lasted only 10-15 days, and ended more gradually than previous events.  The last bubble, that peaked on early April 2015, has not lasted long enough to enable its shape to be discerned with confidence.  Even so, those two events can be adequately modeled by standard bubble functions without plateau, with relatively small error.

Descriptive, not predictive

The model is meant to be purely descriptive, not predictive. That is, it aims only to provide a succint description of the historical prices, without attempting to predict future prices.  If anything, it is "anti-predictive", because it implies that the past bubbles have exponentially decreasing relevance for the future prices, and any future bubbles cannot be detected until they are well underway.

Until mid-2014, it was widely claimed that the price would continue to grow, as it has done in the past, by a sequence of bubbles spaced roughly 9 months apart, with exponentially increasing peak amplitudes. However, this model suggests that such "exponential bubble train model" may be just an illusion.  The good fit of our model to the price, particularly between successive bubbles, suggests that bubbles are added rather than multiplied. It follows that each new bubble is only noticeable when its amplitude is substantially greater than the sum of all tails of the previous bubbles.  If a bubble like the "2011-06" one occurred today, for example, lifting the price by only 10 dollars, it would not be discernible at all.  Therefore, if bubbles actually occurred at random intervals and with random amplitudes, this masking effect would give the impression that bubble amplitudes are increasing.

Interpretation of bubbles as market openings

Although the model is not predictive, I believe that it is explanatory as well as descriptive.  Namely, each bubble can be conjectured to represent a surge in demand due to the opening of some new market.  Each market may be another community of users, isolated from the others by national, language, or legal barriers; or a new use of bitcoins.  

The exponential rise part of a bubble would then be due to the spread of demand in that market by "contagion", possibly amplified by media coverage and speculative demand.  The end of the rise would be due to saturation of that market. (In some bubbles one can see oscillations extending for a month or two after the end of the rise, presumably caused by panic and recovery among the speculators. These oscllations have not been included in the model yet.)  The plateau part of each bubble would be due to a period of relatively constant demand after the peak, while a decreasing tail could be due to gradual decrease of the demand, e.g. for disappointment, government repression, etc.

In particular, the bubble labeled "2013-11" ("Beijing 1", Figure 6) was almost certainly due to the opening of the Mainland Chinese market after the major exchanges Huobi and OKCoin started operating in Beijing, and a report about bitcoin was featured in mainstream Chinese media.  Local reports attribute the huge demand to a large contingent of amateur speculators that used to day-trade in commodities like tea or garlic, and found bitcoin more attractive for that purpose because of its higher volatility.  That is one of the fastest-rising bubbles in the model (about 10%/day).  It peaked at 2013-11-29, when the Central Bank of China (PBoC) intervened and banned the use of bitcoin in e-commerce and forbade financial institutions from dealing with it. That bubble then started decaying at about 0.55%/day.  

The earlier bubble "2013-04" ("Sanghai 1", Figure 5), that has a plateau from about April 5th to May 31, 2013, may have been created by demand in China, too; but by BTC-China in Shanghai, possibly catering to a different community.  BTC-China started operating well before that bubble, but in early 2013 it recruited Bobby Lee as CEO, a Stanford alumnus who formerly worked for Walmart. BTC-China's trading volume started growing exponentially in the first 3 months of 2013, and it was leading the price increase during that period.  In the model, it is assumed that the contribution of that demand to the price remained constant for 2 months and then started to decay.

It would be useful to identiy the markets and/or events that caused the other bubbles.  It has been claimed that one of them was caused by an article in the Wired magazine, for example.

Note that the relation between price increase and demand depends on the liquidity of the market, which is not known, and must be non-linear.  Therefore, the magnitude of each bubble should not be interpreted as a measure of the demand, but only of its effect on the price.

This in interpretation would confirm the claim that the model is not predictive, because the opening of markets in the past would hardly influence the opening of new markets in the future.  In particular, to have another bubble with magnitude above 1000 $/BTC, bitcoin woudl have to conquer some market with demand comparable to the Chinese one.  While there are conjectural candidates for such a market, the past history of the price cannot have much influence on that future event.

Modeling the 2014--2015 price variations

After the all-time high of 2013-11-29, the character of the price graph changed noticeably: instead of typical bubbles, there was a succession of sudden rises and dips, superimposed on the slowly decaying tails of the Chinese bubbles.  Most of these changes were clearly connected to news and rumors that affected either the Chinese or the non-Chinese day traders.  

For example, there was a partial recovery of the price in late December 2013, when some Chinese exchanges regained bank access that had ben closed earlier that month; then a sharp drop on 2014-02-10, then Mark Karpelès of MtGOX cleimed that there was a bug in the protocol; then a recovery on 2014-03-15.  There was also a sharp drop on 2014-03-26, when a Chinese newspaper leaked new of further PBoC restrictions; a recovery on 2014-04-15, when the announced deadline passed with no restrictions; then another sudden drop on 2014-04-25, when the restrictions began to be enforced and one major Chinese exchange had to close.  A sharp but smaller drop occurred on 2014-06-12, when the first USMS auction of the SilkRoad bitcoins was announced, which was reversed on 2014-06-30, when it became known that the auction had been won by entrepreneur Tim Draper.  Other short-lived spikes occurred when it became known that Microsoft was "accepting bitcoins" (which turned out to be only for a few digital products), when it was rumored that McDonalds would start accepting bitcoins on Valentine's day (it was "love" instead), when a statement by an OKCoin staffer was misinterpreted as saying that a large hedge fund would start trading bitcoins, and at several other times.  

These events were quite unlike the previous bubbles in that they had very fast rise and decay (often lasting less than a day), with gradually decaying price in between.  The decay could be attributed to the continuing decay of the main Chinese bubbles "2013-11" and "2014-06"; if these two elements are subtracted from the price, the variations in 2014 and 2015 cn be modeled fairly well by some rectangular pulses with nearly flat tops, and some sharp spikes, all with nearly vertical sides.  These pulses and spikes, in turn can be modeled by degenerate bubble functions with very fast rises and decays.  (Due to their short durations, it is not possible to tell whether the sides are indeed exponentials; but their shape has little effect on the goodness of fit.)  

Figures 7 and 8 below show a an extension of the model of Figures 1 and 2 that uses eight such degenerate bubble functions to model the main events in 2014 and 2015, plus one spike-like element "2011-05" to model the "stutter" during the rise of the "2011-06" bubble.  


[ Figure 7. A 20-bubble model for the price of bitcoin, showing the smoothed actual price (grey), the modeled price (green), and the individual bubbles.  The brown line near the bottom is the ratio of the actual price to the model price.  Click on the image for a full-size version. ]


[ Figure 8. The same data and model in Figure 7, without smoothing. Click on the image for a full-size version. ]

How the model was constructed

A typical bubble has 4 parameters: the rate of change per day during the rise part (always greater than 1), the date of the peak, the rate of change per day during the tail (less than 1, except for the "2010-07" and "2012-08" bubbles), and the magnitude at the peak. Some bubbles had one additional parameter, the duration of the plateau phase.

The number of bubbles in the model (11) and all bubble parameters except the peak magnitude were determined initially by hand, inspecting the price chart. They were then tweaked by variosu means to improve the fitting.

Once the other parameters were chosen, the bubble magnitudes were then determined by weighted least squares fitting of their linear combination to the observed prices P(i).  Each price datum P(i) was assigned a weight W(i) = 1/P(i), in order to simulate the effect of fitting the model in log scale, while actually using linear scale.  

The input prices and the component bubbles were smoothed with a Hann window spanning 2 weeks.

When is the Russian ban coming into effect?  (IIRC, it will be sometime in Q3; if so, that seems unlikely to be the cause...)

Seems a risky business.  Investor buys a batch at 20% premium over the BTC price, but the next day another 100'000 shares mature and their owner puts them for sale, at a 10% discount. 

That may be the reason why there are no real bids yet.  Most of the 1.4 million extant shares are mature. Who knows how many will be put for sale as soon as Greyscale's carrier turtles deliver the certificates...

CYL must be "cut your losses" (or "cut your looses"?) 


It is heretic.

Sorry if this has been reported already: There are still no trades of BIT shares (GBTC) on OTCQX, but the top bids @ 40.00 $/share have been lowered to 38.00 $/share and some other lower price (can't figure out which).

Well, ... except that the thread was started by Adam, and he has been quite active all the time here.  Perhaps he should have a say on what he thinks belongs in the "& discussion" part of the topic?  It not nice to accuse him of negligence, just because his thread is not the way you like...

In 16 months and several thousand posts on this thread, I had only one post deleted by Adam -- not because it was off-topic, but because it was negative about bitcoin and he got upset about it.  The next day, without me asking, he restored the post and apologized.  A fine fellow he is indeed...

Now I suddenly have 10 posts of mine deleted in a row, including the one below, which is quite on-topic:


The last price bubble was China adopting bitcoin; but then China effectively banned its use, so that bubble has been deflating since then.  The next price bubble must come from some demand even bigger than China's.  Russia is about to ban it, India does not seem interested, Africa and Latin America are unlikely to buy much.  Basically the only hope is "Wall Street" and/or IRA accounts.  But that depends on COIN being approved. 

If this is not on-topic for "Price movements & discussion", I don't know what is.

There is now a fairly convincing explanation of what happened, OK.

The reason why people got upset is that the original poster and moderator (@AdamstgBit) never deleted anything for being "off topic"  or "not informative". (I pointed out the size of the thread only to underscore its "anything goes" character, not as an excuse in itself.)  Deletions were limited to actual spam and highly offensive messages.  I myself have reported several such to the Forum moderators.  So, when those atypical deletions of "normal" posts started, people did not know what to think.

Fortress Investment Group (manages 58 billion $ portfolio): bought 20 M$ in bitcoins in late 2013, reported paper loss of 3.7 M$ on that investment in early 2014, invested some money on Pantera's fund management comany.  I can't find any mention of bitcoin by them after that.

Overstock: They were the first big store to accept bitcoin through BitPay. At first they opted to some percentage of their sales in real bitcoins, but then they stopped doing that, they now get 100% dollars.  At the end of 2014 they held only 340,000 $ worth of bitcoins (~1500 BTC), versus 180 M $ in cash and 10 M $ in precious metals(!). Sales paid in bitcoins dropped after the first months, and amounted to only 3 M $ in 2014 (out of ~ 1.5 billion $ in total sales), all domestic.

DigitalBTC: An australian bitcoin mining and trading company now trying to switch to bitcoin software products.  For the 2nd semester of 2014 they reported 1.2 M$ sales of software, 8.7 M$ revenue from bitcoin trading, 4.6 M$ from mining -- but "Net Loss After Tax of $2.3 million due primarily to non-cash accounting adjustments to the fair value of bitcoin inventory and performance rights and depreciation"

Tim Draper: bought 30'000 BTC at the first USMS auction (~650 $/BTC), 2000 at the second one (~350 $/BTC), did not bid for the third.

The Bitcoin Foundation:  nearly broke, had to let go most of its staff; due in part to them keeping their funds in bitcoin.


I don't understand your point.  Yes, Wall Street does not see oil as being worth more than 60 $/barrel right now.  So?

"Something else" in relation to what?  How do you propose to measure the value of oil or bitcoin, except by watching how much people are wlling to bid and ask for it?

I think that the two photos were taken from the same spot too.  For the photo on the right, the camera was placed much lower, so that the view of the building across the street started about 1 floor higher.  

I.e. the bottom row of windows (with 3 panes at top), barely visible in the left picture, is not visible in the right picture.  The middle row of windows in the left photo (narrower, with 2x2 panes) is only half-visible at the bottom of the right photo, and so on.

EDIT: the shift may be 2 floors rather than 1.  The camera may have been lowerd by 0.20 m.  If the window is 1 m away and the other building is 20 m away across the street, then its view would be shifted by 4 m upwards, give or take a few meters.

EDIT 2: given the change in teh camera's position, it is hard to tell whether the sahdows changed much between the two photos.  I would say that the difference is small, less than one hour.  The glare on the other building's windows seems to be about in the same place, which is consistent with that, too.

EDIT 4: in fact, from the right picture we can tell that it was taken from a floor that is at the same height as the floor across the street whose windows are cut in half by the lower windowsill. The windows in that floor have 2x2 panes and are aligned with those of the floor above (2nd row from bottom up in right picture).  The windows on next floor higher up (3rd row up in the right picture) are wider so that the wall slice between them is narrower.  Thus the floor could be identified with certainty from that.

EDIT 5: In other words: the building across the street has the following floors (among others):

A: wide windows with 3x2 panes
B: narrower windows with 2x2 panes, not aligned with teh windows in floor A
C: same as floor B
D: slightly wider windows with 2x2 panes, aligned with the windows of floors C and B.

On the left picture we see a the tops of the windows of floor A, then B, then just a bit of C, on the right side.

On the right picture, we see the top half of the windows of floor B, the windows of floor C, and most of those of D.

Both photos were taken taken from an office on a floor at about the same height as floor B.

As I understand, Gemini (the Winkles' exchange) depends only on ordinary bureaucratic licenses (MSB, MTB, whatever).  If  Coinbase can function, they should be able to function too (unless they are based in NY and have to wait for the BitLicense because of that).

The COIN ETF is more complicated: it must be approved by the SEC, which is not just a bureaucratic formality.  There is no way to tell how long the SEC will take to decide, or whather it will get approved at all.  (In fact, I have read somewhere that, the longer the SEC takes to decide, the less likely is that it will be approved.)

Yes, they are all like that.  Eight short posts of mine now; two comenting on topics often discussed in the thread, the other six about this sudden deletion epidemic.

That thread now has about 240'000 posts (sic) and has traditionally been a rather unrestricted discussion forum, by choice of the OP.  So it must be another admin doing the cleaning on his own initiative.  The deletions may be justified in his view, but are unusual and have left many posters wondering about what is going on, whether it is a hacker or what. 

Perhaps the moderator responsible for the cleanup should post ONE message to the thread about it, so that people would stop wondering and discussing the topic further.

Only that the deletions seem to be random and with no explanation, are affecting all users, and started suddenly a day or two ago...

Random posts are being deleted without explanation from the "Wall Observer" thread and perhaps other places.  The deletion messages have no explanations and say

Wild guess: the user name "Bitcoin_Forum" has admin privileges by default, and someone just created that account.

EDIT: No hacking, just some moderator not aware of the "anything goes" culture of that thread.

That was never a problem: with a trusted centrally managed server network, is trivial to avoid double-spends, it is basic database application. 

The problem that was open for 10-20 years, and that Satoshi hoped to solve with the bitcoin protocol, was avoiding double-spends without a trusted centrally managed network.  That is why bitcoin has miners, proof-of-work, and mining rewards.

Sorry to insist, but I was not trying to say MY opinion of what would be fair, which is totally irrelevant to the situation.  I was trying to guess what KOBAYASHI may decide, by trying to place myself in his place and looking at the situation as HE is probably seeing it, with the constraints that HE may have.  Which should be your priority too.  

Another thing that seems to escape the victims is that Kobayashi is not Mark's replacement.  Mark was the owner of MtGOX, he took your money with certain promises, and therefore he had the obligation to answer your queries promptly, take your demands into account, and make good on his promises.  So it was quite proper to pester him and tell him what you wanted him to do.  But Kobayashi is not the owner of MtGOX, does not owe you anything, and does not have to listen to your wishes.  He works for the Judge, and his obligation is to make the Judge happy by distributing the spoils of MtGOX in a way that the Judge will find fair, respecting all laws and rules.

So, you cannot demand and expect attention from Kobayashi as you were entitled to demand and expect from Mark.  Indeed, the very purpose of bankruptcy protection laws is to stop creditors from banging at the door. The law tells creditors to sit quietly and orderly over there, spell out their claims to the trustee through some standard forms, wait for him to figure out who should get what, and accept the decision once it has been approved by the Judge (subject to legal appeal mechanisms of course).

If you want to improve your chances of getting information from Kobayashi, you should ASK what he intends to do (not tell him what you want him to do); in a formal, polite and humble written letter, not by email or on forums; and preferably a letter written and signed by your lawyer, who will know how to ask meaningful questions with the proper terms.  

If you are about to get a liver transplant, you don't tell the surgeon where you would like to be cut, and you cannot expect him to discuss the details of the procedure with you.  If you have valid concerns, you had better tell them to your doctor, and let him talk to surgeon.  

I don't know, but I guess that Kobayashi deeply regreets having accepted this task.  To begin with, MtGOX was not an ordinary manufacturer or merchant, but a huge ponzi-like scheme; I wonder whether his experience as bankruptcy trustee includes any ponzi.  Claims against a normal company are usually straightforward: "I sent them 20 electric wastebaskets priced at 150 dollars each, here is the invoice, they did not pay me; so I want 3000 dollars".  Claims against a ponzi are like "I deposited 150 dollars, then clicked buttons on their interface hundreds of times, and the interface said that those 150 dollars became 30000 dollars; so I want 30000 dollars".  

Then there is the fact that MtGOX had a huge number of small clients (70'000 minimally active accounts, IIRC) scattered all over the world, who interacted with the company through web interfaces rather than traditional invoices and receipts.  Then there is the fact that MtGOX did not have real accounting for the client balances and trades, just a database operated by buggy and ever-changing programs.  Then Kobayashi found that everything was messed up by a sophisticated and long-running cybercrime that took half a billion dollars from the company's assets, and may or may not have involved complicity of the management...

Kobayashi certainly will not lose money; he will charge suitable fees for his work, subject only to the Judge's  approval.  At least he is doing the parts of the job that must be more familiar to him, such as requesting the return of MtGOX money that Mark loaned to himself, and forcing the bankruptcy of Tibanne, the company that Mark had set up to shield himself from MtGOX's collapse. (Since Mark formally worked for Tibanne who provided manpower services to MtGOX, the money that he made from MtGOX through Tibanne should have been safe from the inevitable bankruptcy of the exchange -- so he may have thought.)  You can see those actions, and the fees paid to Kobayashi and all consultants that he hired, in the reports posted on the MtGOX site.  (Note that each document there has a full and fairly good English translation, at the end of the same PDF file.)

Considering the complexity of the case, the lack of instructions for claim filing, the state of the Tibanne lawsuit, and the fact that the police investigation is still running, I think it is quite likely that Kobayashi will ask the Judge for another extension, which the Judge will probably grant.

You may have seen these reports perhaps?

http://blog.wizsec.jp/2015/04/the-missing-mtgox-bitcoins.html

http://www.ft.com/intl/cms/s/0694b99c-e647-11e4-ab4e-00144feab7de,Authorised=false.html?_i_location=http%3A%2F%2Fwww.ft.com%2Fcms%2Fs%2F0%2F0694b99c-e647-11e4-ab4e-00144feab7de.html%3Fsiteedition%3Dintl&siteedition=intl&_i_referer=#axzz3XlA5R9hf

I can't read the second article because of the paywall. 

Those WizSec guys did a lot of good work, especially their analysis of the blockchain.  However, as they admit, their primary data is the leaked database, which is incomplete and may have been doctored by whoever leaked it.  There is not much that they can do without access to the MTGox wallet files, bank statements, and customer IDs (or even with them, perhaps). 

Imagine someone who wants to find out why a million dollars is missing from a bank account, given only a partial and possibly incorrect list of the deposits and withdrawals, without names.  The strictly honest and obvious answer would be "it cannot be done"; but the guy goes ahead and tries to write up something anyway.  That is how I see their work...

Well, the website does explicitly say that you should not presume that:


See what I mean?  Even though it is written with all letters, you all refuse to even contemplate the idea that YOUR BALANCES ARE NOT WHAT YOU ARE OWED.

And I am not trying to argue what *I* think would be right.  I am only trying to warn you that it is KOBAYASHI who will decide how the claims will be defined; and what *YOU* like, or you think is right, is the least important thing he has to consider.  Laws, precedents, and HIS notion of fairness will come first. 


That is one reason why the claims should not be based on the final balances: in the final weeks, many users felt forced to do bad trades (e.g. selling their coins at 20% of the external market price) because they were led to believe that they could only get yen or dollars out.  Other clients profited from their despair to acquire lots of coins at that ridiculos prices.  The coins have lost value since then, but it is not clear who profited from those "unnatural" trades. 

Even for clients who did not trade in those last few days, their final balances are the resut of trading in an false (probably fraudulent)  market, so their gains or losses were probably affected by fraud (even if unknowingly).  For example, MtGOX's price was significantly higher than the outside market for many months.  It is conjectured that some client was buying coins (probably through the "Willy" robot) with non-existent money; and that the missing coins were already missing by then.  So, profits made by trading inside MtGOX may have been bogus because they were built by trading non-existent coins and money, at prices that were distorted by the fraud.

To me, those distortions make the deposits-minus-withdrawals criterion seems a logical necessity in cases like MtGOX.

Imagine a soccer match were one of the goals happens to be 20% wider than the other, and the diference is discovered only after the game is over.  There is no way to adjust the score after the fact to make it fair: that 20% difference in size may have affected the game in all sorts of ways, not just increased the score by 20%.  The only fair decision would be to cancel that match and play it again.

That's inflation (specifically, consumer price index increase) in March extrapolated for the whole year.  It is still less than the nominal inflation of bitcoin (~9%/year) and MUCH less than the actual change in consumer price index of bitcoin (135%/year in 2014).


Right, but who was complaining about cherry-picking time windows? Since the irresponsible profligate demagogical Keynesian socialists came to power in 2003, inflation has been around 5%/year.  Did I say that it was less than bitcoin's nominal inflation of 9%/year?

But the past performance is not the real problem.  The problem with bitcoin is that it is a ticket to a lottery with unknown prizes and odds. There is no way of even assigning rough probabilities to its value next month being in any range whatsoever.   It may go to zero in a few days, it may go "to the moon", who knows.

What we do know is that the Top Experts in bitcoin have not the foggiest idea of what the price will do, either.  And what I have learned so far tells me that each of the past bubbles were caused by distinct sources of demand that are already exhausted; so the occurrence of a future larger bubble depends on another source of demand with greater magnitude -- which no one can tell whether or when it will appear.

Therefore, selling bitcoin as a hedge against anything, or as good investment, is scamming, pure and simple.  Trying to sell it to the Third World poor, by telling them that it will make them rich, is a crime against humanity.  If you need another million fools to buy your coins and pay for your Lamborghinis, at least look for them among the wealthy, thank you.

It doesn't  look much like one now, does it?  But there are still people trying to sell it as such...


I gather that my previous avatars left some people intimidated and profoundly disturbed by the strength of character that emanated from my majestic beauty.   I was compared to Jimbo Wales, none less.  I hope that you can bear to gaze upon this one without shuddering in awe.