Monthly average USD/bitcoin price & trend

[Ducky1]

Btw, I would like to thank justusranvier https://bitcointalk.org/index.php?action=profile;u=14492 for pointing out the log-log last night on the speculation forum https://bitcointalk.org/index.php?topic=178336.74340

[BitChick]

My alternative approach - log vs log(log) from 01.11.2011 to present. Conclusion: The log-log model seems to better fit the data than the linear-log model, and according to this model we are actually slightly under the trend line. Comments?

Let me get this straight: we've been going super exponential all this time? 

Could you draw this out on a log(log) chart as well?

It may be just a fluke, but super exp. at least fits best for this time period. Here is a log log chart. The blue lines above and below is drawn by hand. The actual price in log-log space have been added manually to the left of the vertical axis. The chart was not really meant for publication so a bit rough.

Thanks a lot for sharing this. Very interesting

BEST CHART EVER!!!!   

I have never been so excited about log charts!

[kdrop22]

My alternative approach - log vs log(log) from 01.11.2011 to present. Conclusion: The log-log model seems to better fit the data than the linear-log model, and according to this model we are actually slightly under the trend line. Comments?

Let me get this straight: we've been going super exponential all this time?  

Could you draw this out on a log(log) chart as well?

It may be just a fluke, but super exp. at least fits best for this time period. Here is a log log chart. The blue lines above and below is drawn by hand. The actual price in log-log space have been added manually to the left right of the vertical axis. The chart was not really meant for publication so a bit rough.

While it is too tempting. Let's not get too carried away by this. Log(Log)) chart implies Bitcoin will be a Million dollar each before end of 2014. We just don't have the infrastructure or adoption for it yet.

[BitChick]

My alternative approach - log vs log(log) from 01.11.2011 to present. Conclusion: The log-log model seems to better fit the data than the linear-log model, and according to this model we are actually slightly under the trend line. Comments?

Let me get this straight: we've been going super exponential all this time?  

Could you draw this out on a log(log) chart as well?

It may be just a fluke, but super exp. at least fits best for this time period. Here is a log log chart. The blue lines above and below is drawn by hand. The actual price in log-log space have been added manually to the left right of the vertical axis. The chart was not really meant for publication so a bit rough.

While it is too tempting. Let's not get too carried away by this. Log(Log)) chart implies Bitcoin will be a Million dollar each before end of 2014. We just don't have the infrastructure or adoption for it yet.

True.  2015 is OK.  That gives everyone a little more time to get the infastructure in place. 

[Roger_Murdock]

So here's an "explanation" for why super-exponential growth might not be completely crazy.

Lots of people have speculated that Bitcoin's adoption might follow the familiar "S-curve" that's been seen with many other new technologies.  And, of course, the initial part of the S-curve looks like exponential growth. And so that's been used to justify the underlying exponential (or so we thought) growth trend for Bitcoin's price that we've witnessed over the past four years.  But why should we assume that the relationship between the rate of adoption and the price is linear?  And, in fact, if the rate of adoption is a measure of the number of users, isn't there at least some reason to expect a non-linear relationship with price? After all, if there are n users in the Bitcoin network, there are n(n-1) / 2 total possible connections.  

kdrop objected that the super-exponential trendline would put us at a million dollars per coin by the end of 2014 which, I have to admit, does seem a tad too optimistic.  But isn't the simple response that the super-exponential trend will only hold true for the initial phase of our now-modified S-curve?  On the other hand, hitting the slowdown / saturation phases in 2014 doesn't feel right to me.

Thoughts?

[BitChick]

So here's an "explanation" for why super-exponential growth might not be completely crazy.

Lots of people have speculated that Bitcoin's adoption might follow the familiar "S-curve" that's been seen with many other new technologies.  And, of course, the initial part of the S-curve looks like exponential growth. And so that's been used to justify the underlying exponential (or so we thought) growth trend for Bitcoin's price that we've witnessed over the past four years.  But why should we assume that the relationship between the rate of adoption and the price is linear?  And, in fact, if the rate of adoption is a measure of the number of users, isn't there at least some reason to expect a non-linear relationship with price? After all, if there are n users in the Bitcoin network, there are n(n-1) / 2 total possible connections.  

kdrop objected that the super-exponential trendline would put us at a million dollars per coin by the end of 2014 which, I have to admit, does seem a tad too optimistic.  But isn't the simple response that the super-exponential trend will only hold true for the initial phase of our now-modified S-curve?  On the other hand, hitting the slowdown / saturation phases in 2014 doesn't feel right to me.

Thoughts?

Maybe the psychological issue of people not wanting to spend so much per coin or fragment of coin could come into play?  I have heard many say that they are "too expensive" etc.  I would hope that next year the more widespread use of mBTC is used which will spur on the super-exponential growth I believe.  But it waits to be seen.

I would think anyone that purchased coins in January of 2013 and saw them grow to almost 100x their investment this year would be a believer that anything can happen!  There is some math to back it up too.  It is not like these numbers are based on just "wishful thinking."   It seems more like a "It' can't be" reaction then "To da Moon" for most people.

[marcus_of_augustus]

It seems more like a "It' can't be" reaction then "To da Moon" for most people.
   

This. The objection is more of an out of hand rejection on grounds of a fundamental disbelief ("It can't be") than a sound, reasoned rejection of the numbers ... the math does not lie, we just have to be sure of the assumptions.

To be fair, I'm still in the "It can't be" camp ... but bad assumptions are more dangerous than bad calculations.

[Voodah]

It seems more like a "It' can't be" reaction then "To da Moon" for most people.
   

This. The objection is more of an out of hand rejection on grounds of a fundamental disbelief ("It can't be") than a sound, reasoned rejection of the numbers ... the math does not lie, we just have to be sure of the assumptions.

To be fair, I'm still in the "It can't be" camp ... but bad assumptions are more dangerous than bad calculations.

But it's not simply that "It can't be", but rather the general observation that infrastructure wise, we're not really there.

Do you really see our current exchanges handling 100x their current load in just 6-7 months from now?

Even if they somehow technologically make it and others pop up to help, how about the banking situation? Bureaucracy takes time. 

[rpietila]

But it's not simply that "It can't be", but rather the general observation that infrastructure wise, we're not really there.

Do you really see our current exchanges handling 100x their current load in just 6-7 months from now?

Even if they somehow technologically make it and others pop up to help, how about the banking situation? Bureaucracy takes time. 

The average trend of 12.4x per year may indeed be the magic number that captures the highest possible growth rate that takes into account all the factors above. It has hardly changed at all during the last 12 months.

My general comment re:the trendlines starting in 11/2011 is that the choice of starting point is arbitrary, which distorts the trend.

[marcus_of_augustus]

It seems more like a "It' can't be" reaction then "To da Moon" for most people.
   

This. The objection is more of an out of hand rejection on grounds of a fundamental disbelief ("It can't be") than a sound, reasoned rejection of the numbers ... the math does not lie, we just have to be sure of the assumptions.

To be fair, I'm still in the "It can't be" camp ... but bad assumptions are more dangerous than bad calculations.

But it's not simply that "It can't be", but rather the general observation that infrastructure wise, we're not really there.

Do you really see our current exchanges handling 100x their current load in just 6-7 months from now?

Even if they somehow technologically make it and others pop up to help, how about the banking situation? Bureaucracy takes time.  

 I agree ... but that doesn't stop the speculative adoption (price) getting way ahead of the utility adoption, and the utility infrastructure that can actually handle that price ....

[Ducky1]

New trading platforms coming online soon-ish http://www.coindesk.com/bitcoin-trading-platform-coinsetter-close-raising-1-5m-new-capital/

[Ducky1]

The average trend of 12.4x per year may indeed be the magic number that captures the highest possible growth rate that takes into account all the factors above. It has hardly changed at all during the last 12 months.

My general comment re:the trendlines starting in 11/2011 is that the choice of starting point is arbitrary, which distorts the trend.

11-2011 is perhaps the start of well working exchanges? I'm to new to really know about this part of history.

[Pruden]

Taking just the data from the last two years is obvious cherry-picking (I have yet to see a less brain-damaged justification than "when exchanges started working properly"). How about taking all data and fitting it to a super-exponential curve? If fitness were better than the exponential trend presented in this thread the theory would be confirmed in my opinion (and the rate of growth would surely be lower than that "$100000 before autumn" nonsense).

[wachtwoord]

Taking just the data from the last two years is obvious cherry-picking (I have yet to see a less brain-damaged justification than "when exchanges started working properly"). How about taking all data and fitting it to a super-exponential curve? If fitness were better than the exponential trend presented in this thread the theory would be confirmed in my opinion (and the rate of growth would surely be lower than that "$100000 before autumn" nonsense).

I would like to see this too. The problem is that very early measuring points are likely to be really inaccurate.

[molecular]

My alternative approach - log vs log(log) from 01.11.2011 to present. Conclusion: The log-log model seems to better fit the data than the linear-log model, and according to this model we are actually slightly under the trend line. Comments?

Let me get this straight: we've been going super exponential all this time? 

Could you draw this out on a log(log) chart as well?

It seems (visually) log(log)) actually does fit better. LOL.

[molecular]

My alternative approach - log vs log(log) from 01.11.2011 to present. Conclusion: The log-log model seems to better fit the data than the linear-log model, and according to this model we are actually slightly under the trend line. Comments?

Wow ... that's some wild-ass result you got!! ... but hard to argue with the math, the fit looks good (just eyeballing but some best-fit stats would be nice) if we are indeed in the super-exponential phase (vertical) right now.

If he's right we've been going super exponential from the start. That's the point.

That's uber-bullish, but can exchanges currently handle that rate of growth?

If each exchange doubles its capacity in time x and the number of exchanges also doubles in time x, it seems possible.

[molecular]

While it is too tempting. Let's not get too carried away by this. Log(Log)) chart implies Bitcoin will be a Million dollar each before end of 2014.

I don't think many here are getting carried away in the sense that we assume much predictive power to this in a serious way. ("It cannot go on forever" certainly applies).

We just don't have the infrastructure or adoption for it yet.

Infrastructure, I agree. Adoption: well, this whole thing is adoption-driven, so adoption is predicted (if you use this model as prediction) by the model itself.

[piramida]

Ducky1, nice graph; I also wondered for most of 2013 why suddenly this radical departure from a steady exponential growth; your log(log) idea is certainly explaining that behaviour. I guess we'll see in 3 months

As for adoption, last month bitcoin has seen more coverage than the previous 4 years. My friends who barely know what Windows is, much less browse internet for anything other than email, already know about bitcoin because they heard about it somewhere, and that was a big surprise for me recently (of course, they don't *understand*, but they've heard - important, too). So 2013 was the year of recognition, 2014 would be the year of adoption.

[rpietila]

Taking just the data from the last two years is obvious cherry-picking (I have yet to see a less brain-damaged justification than "when exchanges started working properly"). How about taking all data and fitting it to a super-exponential curve? If fitness were better than the exponential trend presented in this thread the theory would be confirmed in my opinion (and the rate of growth would surely be lower than that "$100000 before autumn" nonsense).

I would like to see this too. The problem is that very early measuring points are likely to be really inaccurate.

+1. Howbout if somebody older than me would actually take the effort to research into the early days. I was not here then. I think a few hours of actual work (instead of posting) would take us far!

[Ducky1]

Taking just the data from the last two years is obvious cherry-picking (I have yet to see a less brain-damaged justification than "when exchanges started working properly"). How about taking all data and fitting it to a super-exponential curve? If fitness were better than the exponential trend presented in this thread the theory would be confirmed in my opinion (and the rate of growth would surely be lower than that "$100000 before autumn" nonsense).

I only post what I have observed, and what I have observed is in those graphs. I don't know why it works only from 1.11.2011 because I where not following bitcoin at that time, so I have to ask if anything special happened around that time. Its probably possible to fit the whole period, but I suspect it would look rather ugly in the start. Also graphing for prices less than 1$ does not work for log-log unless using complex numbers, so it has to be graphed in cents or lower. Ill do it eventually.