Logarithmic (non-linear) regression - Bitcoin estimated value

[macsga]

According to your chart we should have reached 1000 by now!

But really, if you're right, we've got a long way to recover...

According to the log regression's trend, 1182 is today's estimated value.

You can see it here.

So the price is 76% undervalued.

i am thinking that the 1MB limit is increasingly becoming a dead-weight on the value. This roadblock to scalability is highly publicized now, and it is reasonable that large amounts of money is on the sidelines waiting to see whether Bitcoin can properly scale. The sooner v4 blocks are happening the better the environment will be for the value to return to trend.

Are the debates on this done? There were half people that said yes and the other half said no. Is it decided for sure then?

[1714060802]

1714060802

[1714060802]

1714060802

[1714060802]

1714060802

[1714060802]

1714060802

[uki]

I have a question to the OP: what are the underlying assumptions that make this log regression's trend held over the entire period depicted?
It is hard to believe, taking out some severe doom-like event to the US dollar, to see the price of Bitcoin reaching $100k by 2020. Not even to talk about the valuations beyond the 2020. Isn't it that you try to estimate the trend for much longer time period (next 10, 20, 50 years), having the data from a relatively short period of time (2010 till now) and thus committing a huge prediction error, due to the insufficient amount of data?

[solex]

According to your chart we should have reached 1000 by now!

But really, if you're right, we've got a long way to recover...

According to the log regression's trend, 1182 is today's estimated value.

You can see it here.

So the price is 76% undervalued.

i am thinking that the 1MB limit is increasingly becoming a dead-weight on the value. This roadblock to scalability is highly publicized now, and it is reasonable that large amounts of money is on the sidelines waiting to see whether Bitcoin can properly scale. The sooner v4 blocks are happening the better the environment will be for the value to return to trend.

Are the debates on this done? There were half people that said yes and the other half said no. Is it decided for sure then?

Core dev need to come to agreement on the details, so nothing is "decided", although Gavin has made a proposal which has reasonable support:

60% for, 20% against, 20% whatever

So, only 20% are firmly against changing the limit, which is the same percentage as the polls 2 years ago.

[Trolololo]

I have a question to the OP: what are the underlying assumptions that make this log regression's trend held over the entire period depicted?
It is hard to believe, taking out some severe doom-like event to the US dollar, to see the price of Bitcoin reaching $100k by 2020. Not even to talk about the valuations beyond the 2020. Isn't it that you try to estimate the trend for much longer time period (next 10, 20, 50 years), having the data from a relatively short period of time (2010 till now) and thus committing a huge prediction error, due to the insufficient amount of data?

As said before, this log regression just tries to have a better estimation of bitcoin's value than the linear one. Specially for a 2 or 3 years horizon.

For a longer term value, it's worth to take a look at James d'Angelo's "Bitcoin Price Model".

[Kipsy89]

This is an interesting way to approach the 'real' price of Bitcoin... I've seen many people do similar things. I guess one of the most prominent being rpietila, you know the guy with the castle and the pink car. Haha, man... he's awesome!

[HarmonLi]

I think this is a good approach, although I believe that we may see a behaviour that diverges from the regression line when we hit $10k or $100k...

[uki]

I have a question to the OP: what are the underlying assumptions that make this log regression's trend held over the entire period depicted?
It is hard to believe, taking out some severe doom-like event to the US dollar, to see the price of Bitcoin reaching $100k by 2020. Not even to talk about the valuations beyond the 2020. Isn't it that you try to estimate the trend for much longer time period (next 10, 20, 50 years), having the data from a relatively short period of time (2010 till now) and thus committing a huge prediction error, due to the insufficient amount of data?

As said before, this log regression just tries to have a better estimation of bitcoin's value than the linear one. Specially for a 2 or 3 years horizon.

For a longer term value, it's worth to take a look at James d'Angelo's "Bitcoin Price Model".

I know that this is your time window, but still my question that remains unanswered is: do you have enough data so far, to make such regression reliable?
So far you had like 2081% above the line and 81% to the downside, which indicates that the fit to the curve is far from acceptable in terms of prediction error. Meaning probably you don't have enough 'training data' to make any reliable predictions for the future.
Just to illustrate what I am talking about, try a small exercise: take only first two years of data and make the regression based on that. Compare that to the real results in terms of errors. You will probably see even worse results in terms of how far it is from reality.

[mmortal03]

I have a question to the OP: what are the underlying assumptions that make this log regression's trend held over the entire period depicted?
It is hard to believe, taking out some severe doom-like event to the US dollar, to see the price of Bitcoin reaching $100k by 2020. Not even to talk about the valuations beyond the 2020. Isn't it that you try to estimate the trend for much longer time period (next 10, 20, 50 years), having the data from a relatively short period of time (2010 till now) and thus committing a huge prediction error, due to the insufficient amount of data?

As said before, this log regression just tries to have a better estimation of bitcoin's value than the linear one. Specially for a 2 or 3 years horizon.

For a longer term value, it's worth to take a look at James d'Angelo's "Bitcoin Price Model".

I know that this is your time window, but still my question that remains unanswered is: do you have enough data so far, to make such regression reliable?
So far you had like 2081% above the line and 81% to the downside, which indicates that the fit to the curve is far from acceptable in terms of prediction error. Meaning probably you don't have enough 'training data' to make any reliable predictions for the future.
Just to illustrate what I am talking about, try a small exercise: take only first two years of data and make the regression based on that. Compare that to the real results in terms of errors. You will probably see even worse results in terms of how far it is from reality.

Unless you're willing to propose some sort of non-standard regression function that makes sense as applied to the Bitcoin market, complaints about 2081% and -81% are unfounded here. The fact that the Bitcoin market has volatility with respect to the regression trend does not mean that this chart offers nothing worthwhile in terms of representing reality. To get an equation to follow the price more closely, you'd have to throw in something more exotic, like an oscillation component or something, and doing so just to get a better fit is often not the best tactic.

[thejaytiesto]

This is an interesting way to approach the 'real' price of Bitcoin... I've seen many people do similar things. I guess one of the most prominent being rpietila, you know the guy with the castle and the pink car. Haha, man... he's awesome!

He predicted 1 Million per coin back then. I wonder where does he stand nowadays and if he still believes that will happen and when.

[cbeast]

This is an interesting way to approach the 'real' price of Bitcoin... I've seen many people do similar things. I guess one of the most prominent being rpietila, you know the guy with the castle and the pink car. Haha, man... he's awesome!

He predicted 1 Million per coin back then. I wonder where does he stand nowadays and if he still believes that will happen and when.

He claims he dumped a lot of his bitcoins for monero. He may have been pumping and dumping.

[uki]

I know that this is your time window, but still my question that remains unanswered is: do you have enough data so far, to make such regression reliable?
So far you had like 2081% above the line and 81% to the downside, which indicates that the fit to the curve is far from acceptable in terms of prediction error. Meaning probably you don't have enough 'training data' to make any reliable predictions for the future.
Just to illustrate what I am talking about, try a small exercise: take only first two years of data and make the regression based on that. Compare that to the real results in terms of errors. You will probably see even worse results in terms of how far it is from reality.

Unless you're willing to propose some sort of non-standard regression function that makes sense as applied to the Bitcoin market, complaints about 2081% and -81% are unfounded here. The fact that the Bitcoin market has volatility with respect to the regression trend does not mean that this chart offers nothing worthwhile in terms of representing reality. To get an equation to follow the price more closely, you'd have to throw in something more exotic, like an oscillation component or something, and doing so just to get a better fit is often not the best tactic.

I understand that the idea is to have something very simple and non-complex to approximate the price of Bitcoin, but still I feel uneasy (even very uneasy) when I see the prediction error and the number of data samples on which the trend is based. And I believe that is perfectly founded complain, to which I haven't got any response yet, rather than that one regression is better than the other (linear one). What if both are not closely enough approximating the trend?
Now, you also say that it makes a perfect sense to use such a regression. Can you explain in more detail, what backs that statement?

[pereira4]

I have a question to the OP: what are the underlying assumptions that make this log regression's trend held over the entire period depicted?
It is hard to believe, taking out some severe doom-like event to the US dollar, to see the price of Bitcoin reaching $100k by 2020. Not even to talk about the valuations beyond the 2020. Isn't it that you try to estimate the trend for much longer time period (next 10, 20, 50 years), having the data from a relatively short period of time (2010 till now) and thus committing a huge prediction error, due to the insufficient amount of data?

Who would have thought in 2009 that bitcoin could reach 1 per coin? we have the same mentality towards 10 or even 100K. Everything impossible in the unknown future.

[mmortal03]

Unless you're willing to propose some sort of non-standard regression function that makes sense as applied to the Bitcoin market, complaints about 2081% and -81% are unfounded here. The fact that the Bitcoin market has volatility with respect to the regression trend does not mean that this chart offers nothing worthwhile in terms of representing reality. To get an equation to follow the price more closely, you'd have to throw in something more exotic, like an oscillation component or something, and doing so just to get a better fit is often not the best tactic.

I understand that the idea is to have something very simple and non-complex to approximate the price of Bitcoin, but still I feel uneasy (even very uneasy) when I see the prediction error and the number of data samples on which the trend is based. And I believe that is perfectly founded complain, to which I haven't got any response yet, rather than that one regression is better than the other (linear one). What if both are not closely enough approximating the trend?
Now, you also say that it makes a perfect sense to use such a regression. Can you explain in more detail, what backs that statement?

I don't think more samples would make it more accurate for forecasting. This isn't something where we are lacking data points. What we've got is the best that we can have. You could use intra-day data for more samples, for example, but that wouldn't help anything; the daily weighted price is fundamentally representative of that.

Assumptions have to be made. Trolololo has made some. It's likely that the information necessary to best predict the price may not solely be contained in the past price data. Keep in mind, price growth can't just continue forever. But determining what outside parameters may predict where the model will diverge from a particular pattern is a hard question.

Read through the following for some formal assumptions and interpretation: https://en.wikipedia.org/wiki/Linear_regression#Assumptions

Also, the following:

If the guys making the charts could actually predict the market movements, they'd be rich

Edit:

The following is also helpful: https://en.wikipedia.org/wiki/Overfitting

[uki]

Unless you're willing to propose some sort of non-standard regression function that makes sense as applied to the Bitcoin market, complaints about 2081% and -81% are unfounded here. The fact that the Bitcoin market has volatility with respect to the regression trend does not mean that this chart offers nothing worthwhile in terms of representing reality. To get an equation to follow the price more closely, you'd have to throw in something more exotic, like an oscillation component or something, and doing so just to get a better fit is often not the best tactic.

I understand that the idea is to have something very simple and non-complex to approximate the price of Bitcoin, but still I feel uneasy (even very uneasy) when I see the prediction error and the number of data samples on which the trend is based. And I believe that is perfectly founded complain, to which I haven't got any response yet, rather than that one regression is better than the other (linear one). What if both are not closely enough approximating the trend?
Now, you also say that it makes a perfect sense to use such a regression. Can you explain in more detail, what backs that statement?

I don't think more samples would make it more accurate for forecasting. This isn't something where we are lacking data points. What we've got is the best that we can have. You could use intra-day data for more samples, for example, but that wouldn't help anything; the daily weighted price is fundamentally representative of that.

Assumptions have to be made. Trolololo has made some. It's likely that the information necessary to best predict the price may not solely be contained in the past price data. Keep in mind, price growth can't just continue forever. But determining what outside parameters may predict where the model will diverge from a particular pattern is a hard question.

Read through the following for some formal assumptions and interpretation: https://en.wikipedia.org/wiki/Linear_regression#Assumptions

Also, the following:

If the guys making the charts could actually predict the market movements, they'd be rich

Edit:

The following is also helpful: https://en.wikipedia.org/wiki/Overfitting

As I wrote in one of my previous posts, I am not expecting a perfect fit here, but the assumption on the continuously growing trend that is made by Trolololo is somehow doubtful. And that is why I made my comment of having not enough data (too short observation window). To illustrate what I mean by having not enough data, just a little thought experiment: how would look the regression curve in comparison to the present one in, say year 2020, if hypothetically we would stay in the down trend till then? Or another, more practical life experiment: you throw a stone upwards and plot its height as it goes up (say you sample every 1 millisecond). If you take the regression on the very first few seconds before the stone starts falling, you would probably see a very different curve, than what you would plot if you have a longer time window.

[mmortal03]

If you take the regression on the very first few seconds before the stone starts falling, you would probably see a very different curve, than what you would plot if you have a longer time window.

Of course. Didn't I post a link to overfitting? The following is also the common reply: http://xkcd.com/605/

Anyway, this isn't a lack of data issue. It's more of an issue of having a theory. Trololo has a theory of what will happen, and this is that it will keep increasing along this curve, on average, to at least 2024. What is your theory? What is the evidence that you have that might result in a different curve than what Trololo is theorizing? It's one thing to criticize, it's another to put your money on your own curve. If you were to draw one, what would it look like, and what factors would this be based n?

[usairx]

I personally think Bitcoin is undervalued in the long term, but I'm very curious and interested to see what happen in 2015

[HarmonLi]

Recently I really fell in love with the charts suggesting that both 2013 bubbles were artificially created (or at least inflated, rather) by Mt. Gox. If you take those bubbles out of the equation you get a nice long-term support we're bouncing off right now. Nice.

[mmortal03]

This is an interesting way to approach the 'real' price of Bitcoin... I've seen many people do similar things. I guess one of the most prominent being rpietila, you know the guy with the castle and the pink car. Haha, man... he's awesome!

He predicted 1 Million per coin back then. I wonder where does he stand nowadays and if he still believes that will happen and when.

He claims he dumped a lot of his bitcoins for monero. He may have been pumping and dumping.

He did, but he's only pumping Monero, not dumping it.

[Trolololo]

Since some days ago we are in record undervaluation: -81%

[Natalia_AnatolioPAMM]

I personally think Bitcoin is undervalued in the long term, but I'm very curious and interested to see what happen in 2015

its going to crash down to 50. i strongly believe that