Back in 2011, there was great discussion in this forum about maximum bitcoin value. Once you wrap your mind around the notion that bitcoin could disrupt precious metals as a store of value, and disrupt fiat currency, i.e. M2 monetary base as a payment mechanism, then you begin to accept that $1,000,000 per bitcoin is a candidate ultimate value for the virtual crypocurrency.

But given a wildly high maximum price - such as $1,000,000 per bitcoin, then how do we get there? ... well by exponential growth. It was one thing to watch bitcoin go from $1 to $30 back in 2011 but quite another to watch it go from $65 to over $400 in the last few months. The larger absolute numbers, as well as the impact on our unrealized profit - are staggering.

First, thanks for the welcome. I am drawn to this forum, like a moth to a lighted window, whenever there is a bitcoin bubble. I am through with trading and just buying. Plus I have an affection for Risto viewpoints.

Regarding the trendline connecting the bottoms of the previous two bubbles, I think now that 2012 was a period of consolidation, where the uncertainty of bitcoin held back speculation - that was released in January-April 2013. It now appears that 2012 was not the trend, rather it was below the trend.

In accordance with the conventional wisdom that bitcoin prices will be marked by a series of bubbles before settling down as a mature financial instrument, I believe that this bubble and future bubbles will have the form of damped oscillations rather than the classically symmetrical financial bubble pattern exhibited by the June 2011 peak.

Can you restate your issue about the S-curve? I plotted it log 10 so the exponential part is linear on the chart.

As a matter of fact, I met Eric Raymond in Austin where I worked at an AI research company. His comments on democratic politics are interesting. But I am satisfied that economic injustice is solvable by the political process. Not absolutely solvable - just good-enough solvable. In particular, I believe that household purchasing power, when unfairly allocated, can be recognized and fixed by the political process, stemming from the will of the voting majority in democratic jurisdictions, or by well-meaning technocrats elsewhere.

Suppose in some possible world, that the three Trinity College crypto guys are revealed to be Satoshi Nakamoto, and suppose that in 2018 they are bitcoin trillionaires. Would not the UK do something about redistributing their windfall? Or more likely, they would voluntarily distribute the majority of it.

Regarding a defense of the Technological Singularity, it is off-topic here. Suffice it to say that I believe all relevant intelligent human behavior, and likewise the behavior of human organizations, can be modeled as learnable skills. My own research, following the advice of Alan Turing, is to create a computer system capable of being taught skills, and then proceed to teach it.

Your blog post mentions human creativity as a counterexample. An important AI skill that I am developing is the ability for the computer to write and modify computer programs. This is the sort of creativity that interests me.

Potential bitcoin profit provides yet another way to bootstrap this effort. And I have in mind using bitcoin as an internal currency for the system of human-mentored software agents that comprise my AI architecture.

To be precise, in a logistic function, i.e. sigmoid, the acceleration slows from the starting point to the mid point. In the historical price series, the slowing is not yet noticeable.

My chart employs a hand-fit logistic curve that passes though the beginning and ending points of the historical price series that was convenient for me to use. My analysis neglects, or rather smooths out, technically relevant highs and lows that characterize the tradingview chart. Also my price history is more recent.

The logistic analysis was mainly to figure out the circumstances that will end the exponential growth of bitcoin prices. I do not believe those results are any better than conventional technical analysis of where prices might be next year.

What is very interesting about the tradingview chart, and its "Mother of all triangles", is the technical prediction it makes about the current bubble. It looks like we are approaching the upper resistance line of the triangle at about $550, using MtGox prices.

Was running Lubuntu, but now running vanilla Ubuntu on development box and server.

When I occasionally need Windows 7, I launch it in VirtualBox.

Two years ago I ran three barebone miners on Mint.

I also have an Mac mini at my vacation home that I needed last year for iOS development. While there I usually launch Ubuntu, again from VirtualBox.

The market shook off yesterday's 10% correction, and while watching it I seemed to me that neither are the buyers exhausted yet, nor are the large sellers successful in signalling each other that this is the big one, i.e. the collapse.

I think we go much higher and faster before a collapse. But I am buying more today regardless because I have faith that prices in 2014 will average well over $1,000.

You should be able to find all the talent and collaborators you need on the Alternate cryptocurrencies section of the forum. Metaphorically, you are pushing a boulder uphill here whereas over there your boulder rolls itself downhill. Best wishes.

Here is an article ... http://www.coindesk.com/bitcoin-exchange-btc-china-eliminates-trading-fees/

I recall somewhere that BTC China still makes money with a fiat withdrawal fee.

I agree that this does not yet have the feel of a big correction, where bears sort of signal each other with consecutive sales, following an exhaustion of the buyers  at the peak.

Gox charges about half of one percent per trade, so a round trip costs about 1 percent of the timiing-motivated trade. Furthermore, the bid/ask spread also subtracts from round-trip profit. Even a good mechanical trading system such as trend-following moving averages is more about protecting capital from downdrafts than beating buy-and-hold. Lucky those folks trading on BTC China with a temporary suspension of trading fees.

After surrendering a substantial amount of my coins to gox fees, I am through with timing and resigned to buy more immediately and hold in an offline paper wallet for at least 48 months.

China is leading the way down from 2700 back to test resistance at 2500.

I see a steady drift downwards over the last four hours towards the resistance level near $390. Volume is higher at BTC China, so perhaps the sell off will continue for awhile.

There are some great threads on this forum that others may guide you to. But here is a video tutorial ... http://www.youtube.com/watch?v=JsyPfiENwQU

Again, best wishes.

I accept and understand your well-founded critique of my use of the logistic function to model bitcoin price behavior.

Regarding the ability to short bitcoin, I participated last spring in the lending program provided by Hong Kong-based Bitfinex, who provide margin and short selling for bitcoin speculators. I withdrew from the modest bitcoin returns that I gained due to my perceived risk, remote but nonetheless catostrophic, that the exchange could be hacked or bankrupted. In addition to the well established Bitfinex, a just-announced startup Coinsetter in New York says that they will be fully compliant with applicable US financial institution regulations and offer short selling too. You can purchase short options on MPEx http://mpex.co/ .
  

This is a good implementation of the classic buy-and-hold strategy.

Yes you will be rich    - most of the appreciation will occur in the next 48 months according to my logistic analysis of bitcoin prices. Put 40 BTC into dukong's ranking calculator to see where you would stand with regard to other holders.

http://btc.ondn.net/search

Please take security seriously. I created a paper wallet from bitaddress.org and put it in my safe deposit box. Then you can forget about it.

Thanks for your thoughtful comments.

I would say that at the midpoint, half of the population of convincible speculators have been convinced to buy. Rpietila has questioned just how much capital, i.e. fiat-to-bitcoin purchases are required to move the price at any given point. We need to think more about this aspect. But my intuition is that because the price is still higher past the mid-point, that the total capital input is still greater past the mid-point even as fewer speculators contribute it. My logistic model rationale is that the constraint on bitcoin price growth is that committed capital is exhausted beyond some maximum price.


Yes, indeed these are the assumptions of the model and you illustrate its deficiencies. However, the curves I fitted demonstrate that the exponential growth, even in the most wildly bullish scenario of a $5,000,000 maximum bitcoin price, stops exponentially growing within 48 months. Because bitcoin price appreciation front-runs the underlying bitcoin economy, there will be some set of expectation-changing circumstances that end the exponential growth and mark the mid-point of speculation adoption. The factors such as those you mention might just be those expectation-changing events.

Rpietila believes that there will be a final blow-off crash at the peak bitcoin price. You may believe that bitcoin may crash way back down, but I think that the final blow-off peak will be resolved by a damped oscillation similar in shape to April-August 2013 - settling down to a bumpy plateau which will benefit the underlying bitcoin payment economy as it disrupts the current payment infrastructure.

Shaped like this ...

I think that I understand your concern now. Obviously you agree now that an ideal logistic model does not exponentially increase without end, rather after the mid point of adoption, there is an exponential decrease in the adoption rate ...



My model attempts to fit a logistic curve to the price history of bitcoin. Stretch your mind a bit. We agree that exponential growth cannot go on forever, but how does it end for bitcoin prices?

Rather than  model the population of bitcoin technology innovation adopters, I am modelling the population of bitcoin financial speculators who transfer fiat to bitcoin. My rationale is that motivation to buy and hold bitcoins spreads through the finite population of speculators in a manner which positively depends on the size of the already-convinced speculator population, and is limited by the amount of funds that speculators are willing to transfer from fiat to bitcoin at a given bitcoin price. The logistic model has only two parameters, the adoption period (X axis above) and the maximum population size (Y axis above). I guess at the latter, i.e. what is the maximum price that speculators will pay when every convincible speculator has been convinced to buy. And I determine the adoption period by fitting the curve to the historical bitcoin price data.

Because the bitcoin price series is subject to periodic bubbles and crashes, e.g. most recently in the form of dampened oscillations, my model only suggests a likely trend for any supposed maximum bitcoin price.

Here is the shared spreadsheet that contains the model, which can be viewed - or copied and edited ... https://docs.google.com/spreadsheet/ccc?key=0ArD8rjI3DD1WdGhDN3FBWFptTlZTREN0cFkxZ3JHTnc

Great!

My university degree is in applied mathematics. Please explain with what math you learned why a logistic function is inherently unstable, as no exponential growth goes on forever.

First, I do not appreciate your hijack of this important thread.

Second, my post is about using a logistic function to model bitcoin prices. Perhaps your reasoning is shallow, e.g. drawing wrong conclusions at a glance. There is an excellent Wikipedia article that explains what a logistic function is - and explains its many uses.

http://en.wikipedia.org/wiki/Logistic_function

Seriously, if you do not understand the math I would be glad to teach it to you.

My logistic analysis is akin to your own log regression analysis. It does not say much about the bubble fluctuations that we have seen in the past and expect in the future. Rather it suggests some average trends.

And could not agree more with your timing advice.