BETI: Bitcoin Exponential Trend Index and technical analysis

[jehst]

It's self-correcting. No need to change anything.

[1714487834]

1714487834

[1714487834]

1714487834

[1714487834]

1714487834

[1714487834]

1714487834

[1714487834]

1714487834

[uki]

It's self-correcting. No need to change anything.

Yes, I know it is self correcting, and I like this improvement. My next question is thus: what is the time constant for this self correction? Choice of this parameter will have the crucial impact on the expected value and thus the accuracy of the model.  

[afbitcoins]

Very interesting to see how low we are on this index which I take to mean there is huge potential upside if log increase is to continue.

I'm not totally clear how this is worked out though. Does this sound about right? It takes historical data and calculates a mean exponential rise based on all data points accumulated to date. Then shows how far above or below that mean we are currently ?

sumthin' like that ?

[marcus_of_augustus]

Interesting ... can't decide if fun or substantial yet.

[afbitcoins]

It looks a bit like a giant RSI indicator. or is it just me?

[rpietila]

It's self-correcting. No need to change anything.

Yes, I know it is self correcting, and I like this improvement. My next question is thus: what is the time constant for this self correction? Choice of this parameter will have the crucial impact on the expected value and thus the accuracy of the model.  

There is no time constraint. It will always take all data into account. The only possible error is if the "exponential fit" is not the correct fit for the trendline. For a long time it was (giving the highest R^2) but now it is not sure. I haven't tested for some time. It is possible that a "inverted parabola in the exponential chart" is giving you a better fit (higher R^2 value). If it was so, Bitcoin would have been a fad that is destined to collapse in the next 2 years totally.

In the end, you need to deduce yourself from the fundamentals if Bitcoin is here to stay or not. If you think it is, it is probably a given that the current price is very favorable for buying (unless you are already very heavily invested). If not, this is a very good time to exit.

[uki]

It's self-correcting. No need to change anything.

Yes, I know it is self correcting, and I like this improvement. My next question is thus: what is the time constant for this self correction? Choice of this parameter will have the crucial impact on the expected value and thus the accuracy of the model.  

There is no time constraint. It will always take all data into account. The only possible error is if the "exponential fit" is not the correct fit for the trendline. For a long time it was (giving the highest R^2) but now it is not sure. I haven't tested for some time. It is possible that a "inverted parabola in the exponential chart" is giving you a better fit (higher R^2 value). If it was so, Bitcoin would have been a fad that is destined to collapse in the next 2 years totally.

In the end, you need to deduce yourself from the fundamentals if Bitcoin is here to stay or not. If you think it is, it is probably a given that the current price is very favorable for buying (unless you are already very heavily invested). If not, this is a very good time to exit.

I was referring to the time constant, not constraint, that is a typical parameter describing how quickly you are converging to the asymptotic value: https://en.wikipedia.org/wiki/Time_constant. That is probably included in this self-correction property that updates the initial algorithm, as far as I understood it.

As for time limitation of the data that is taken into account, that was my previous comment, that probably the initial data (before the first peak at $35) may be misleading for the exponential trend. But I was answered that this is adjusted with that self-correcting property.

[DieJohnny]

It's self-correcting. No need to change anything.

Yes, I know it is self correcting, and I like this improvement. My next question is thus: what is the time constant for this self correction? Choice of this parameter will have the crucial impact on the expected value and thus the accuracy of the model.  

There is no time constraint. It will always take all data into account. The only possible error is if the "exponential fit" is not the correct fit for the trendline. For a long time it was (giving the highest R^2) but now it is not sure. I haven't tested for some time. It is possible that a "inverted parabola in the exponential chart" is giving you a better fit (higher R^2 value). If it was so, Bitcoin would have been a fad that is destined to collapse in the next 2 years totally.

In the end, you need to deduce yourself from the fundamentals if Bitcoin is here to stay or not. If you think it is, it is probably a given that the current price is very favorable for buying (unless you are already very heavily invested). If not, this is a very good time to exit.

WTF rpietila, a good time to exit..... it is too late to exit! Bitcoin replaced by what? Attacked by what? Destroyed by what?

[rpietila]

I was referring to the time constant, not constraint, that is a typical parameter describing how quickly you are converging to the asymptotic value: https://en.wikipedia.org/wiki/Time_constant. That is probably included in this self-correction property that updates the initial algorithm, as far as I understood it.

I thought the model was simply trying to find the best-fitting (least squares method) linear trendline for all the price data in the logarithmic domain. With the transformation to non-logarithmic, it will be the same as the best-fitting exponential trendline in the linear domain.

I know a little math, but not in English. Therefore it may sound funny, but I am pretty sure no time constants are used, just recalculation of the trendline.

[rogerwilco]

Bitcoin replaced by what? Attacked by what? Destroyed by what?

Monero, Obama, you touching yourself at night.

[jl2012]

It's self-correcting. No need to change anything.

Yes, I know it is self correcting, and I like this improvement. My next question is thus: what is the time constant for this self correction? Choice of this parameter will have the crucial impact on the expected value and thus the accuracy of the model.  

There is no time constraint. It will always take all data into account. The only possible error is if the "exponential fit" is not the correct fit for the trendline. For a long time it was (giving the highest R^2) but now it is not sure. I haven't tested for some time. It is possible that a "inverted parabola in the exponential chart" is giving you a better fit (higher R^2 value). If it was so, Bitcoin would have been a fad that is destined to collapse in the next 2 years totally.

In the end, you need to deduce yourself from the fundamentals if Bitcoin is here to stay or not. If you think it is, it is probably a given that the current price is very favorable for buying (unless you are already very heavily invested). If not, this is a very good time to exit.

Rsq of the linear trend line is only 0.432, while the exponential trend line is 0.82

[0x3d]

So how does that last drop look on BETI? Care to update yet, jl? Thx.

[DieJohnny]

If we are in a bubble, even a mini one, how high it may go? The Bitcoin Exponential Trend Index (BETI) may give us some hints. The mid-2012 mini bubble went from -1.25 to -0.5 (+0.75). The early-2013 bubble went from -0.75 to 1.5 (+2.25). The late-2013 bubble went from -0.125 to 1.625 (+1.75).

After the long bear trend, I believe we are more likely in a mini bubble like the mid-2012 one, that may send us from -2 to -1.25 (+0.75) in a few weeks. In that case, the target price will be around $530 for this mini bubble. A full bubble may happen next year around halving which will create a new ATH and push BETI above 0.

In my opinion a full bubble will likely follow in less than three months but no sooner than one month after the mini-bubble.

The full bubble might be on a scale we haven't seen before.

This might be wishful thinking, but most of the charts seem positive on this.

I don't think it is wishful thinking. Compare today with Dec 2013, who was speculating and ABLE to buy bitcoin in 2013 compared to today. It makes completely rational sense that the top will be vastly higher than in 2013 even with Gox fraud willy bot.

Can we get some updates on this thread? I think we are tippping.

[jl2012]

Sorry for no update for months:

Date:    21-Feb-2016
BETI   -1.546
VWAP:    438.18
x:    2045
a:    0.00406
b:    -0.67764
Rsq:    0.81243
The day's expected price:    2055.71
Actual price / expected price:   21.32%
Price to break the -2.23 all-time-low   221.32
Price to break the +1.87 all-time-high   13308.40
Predicted date for today's price:    5-Feb-2015
Days ahead:    -380.57
Daily price rank:    314
   
   
(See OP for explanation)   
   
   
   
https://www.wolframalpha.com/input/?i=e+%5E+%28+0.00406162298379092++%28+number+of+days+since+jul+17%2C+2010+%2Fdays+%29+-0.67764128572232+%29   

[jl2012]

[jl2012]

I am recently focusing on the development of segwit, and also the HF drama.

Seems we built a nice bottom at -1.7 with a ceiling at about -1.42

[birr]

Are you going to try fitting a logistic?

[jl2012]

Date:    10-Jun-2016
BETI:          -1.371
VWAP:    576.10
x:    2155
a:    0.00382
b:    -0.51026
Rsq:    0.80310
The day's expected price:    2269.70
Actual price / expected price:   25.38%
Price to break the -2.23 all-time-low   244.36
Price to break the +1.87 all-time-high   14693.72
Predicted date for today's price:    17-Jun-2015
Days ahead:    -358.69
Daily price rank:    192
   
   
(See OP for explanation)   
   
   
   
https://www.wolframalpha.com/input/?i=e+%5E+%28+0.00382257957102623++%28+number+of+days+since+jul+17%2C+2010+%2Fdays+%29+-0.510256735157911+%29   

[jl2012]

It seems the resistance at -1.4 has been broken.

[ssmc2]

Thanks for the update! Always liked this thread. So do you think we'll encounter more resistance at -1.2?