BETI: Bitcoin Exponential Trend Index and technical analysis

[Bitcoinaire]

It seems more likely that bitcoin is on a logarithmic regression curve, like so:

The chart has been updated recently by Marc De Mesel
Below is the link to his YT video.
I refer to this chart and believe would could see more than 75% undervaluation

https://www.youtube.com/watch?v=k9JsmelO12s

He's counting 2015 as the beginning of the bull market, but it was range bound ($200-$300) the first 10/11 months.

[1713527867]

1713527867

[1713527867]

1713527867

[1713527867]

1713527867

[birr]

The range is moving down via a SQRT curve of sorts - we aren't ever breaking zero again and if we touch -.5, sell everything! I like this indicator, but it should somehow evolve away from log and towards a more gradual curve. I patiently waited for another 1.5 on the last run up without realizing this.

No need to abandon the log scale.  Only needs judicious choice of curve.  SQRT it is.
I ran some numbers on the price history and fitted a square root baseline.  Vertical axis, log of dollars.



This chart shows the relation between the price and the baseline over time.

[of00160]

Just a quick update...

Date   11/28/18
VWAP   $4,098.94
x   3056
y   8.318484
a (slope) historical   +0.0032121222
b (intercept) historical   -0.0144982009
Rsq historical   0.869245
BETI historical   -1.483263
Expected price historical   $18,065.28
Actual price / expected price historical   22.68% (kek!)
 
-2.00   $2,444.87
-1.50   $4,030.91
-1.00   $6,645.84
-0.50  $10,957.15
-0.25  $14,069.25
+0.00  $18,065.28
+0.25  $23,196.27
+0.50  $29,784.61
+1.00  $49,106.52

The BIT historical is BElow -2 as of 1/22/19. The market is seriously undervalued.

[keystroke]

-2.15 at the moment.

[Bitcoinaire]

It looks like we have to break -1.75 to be sure we've left the bear market and are in a new bull market.

[of00160]

It looks like we have to break -1.75 to be sure we've left the bear market and are in a new bull market.

That could like breaking above 6k.
The trololo line on Tradingview is at fair value.
Btc have be trading on avg above that line the last 7years.
The trololo line is at ~3500

[Biodom]

I am getting BETI at -2.083 currently using the formula described here:

https://bitcointalk.org/index.php?topic=470453.msg26854986#msg26854986

Since -2.25 is an all time low (in 2011) with second low in 2015 at -2.2, next point in progression might indeed be -2.15.
Perhaps, -2.15 would be IT for this cycle.

Next ATH might not even get to BETI=1.0 (achieved in 2017) and stop somewhere lower if peaks are actually on a declining trend (too small number of points so far).
Last peak was somewhat disappointing in that regard as I expected at least a match to 2013.

[Dogboy714]

I am getting BETI at -2.083 currently using the formula described here:

https://bitcointalk.org/index.php?topic=470453.msg26854986#msg26854986

Since -2.25 is an all time low (in 2011) with second low in 2015 at -2.2, next point in progression might indeed be -2.15.
Perhaps, -2.15 would be IT for this cycle.

Next ATH might not even get to BETI=1.0 (achieved in 2017) and stop somewhere lower if peaks are actually on a declining trend (too small number of points so far).
Last peak was somewhat disappointing in that regard as I expected at least a match to 2013.

We will never break 0 again, not to mention 1.0.

[exstasie]

We will never break 0 again, not to mention 1.0.

Don't get too comfortable with the current regression model, based on such limited data. We cant assume based on extrapolation that Bitcoin will never rise above the upper bounds of the historical trend. That would eliminate the possibility of an S-curve model for adoption, which I believe is still possible.

[Dogboy714]

We will never break 0 again, not to mention 1.0.

Don't get too comfortable with the current regression model, based on such limited data. We cant assume based on extrapolation that Bitcoin will never rise above the upper bounds of the historical trend. That would eliminate the possibility of an S-curve model for adoption, which I believe is still possible.

Mmmm...hopium! Not going to happen, though. Dollar cost average in now and HODL til $80k - then sell everything and retire! Don't forget the last part about selling everything - the hype will be stronger than you could possibly imagine but you've gotta go contrarian and sell. Rinse and repeat for more bitcoin if you like, but get out!

[josegines]

and HODL til $80k -

why $80k?

[Biodom]

and HODL til $80k -

why $80k?

right now BETI is about -2.09.
About 29K at BETI=0
80K at BETI=1.0 (as the last peak).

However, this is just for today.
80K number might have no meaning (not close to any round number) in a year or two.

[Wilhelm]

and HODL til $80k -

why $80k?

right now BETI is about -2.09.
About 29K at BETI=0
80K at BETI=1.0 (as the last peak).

However, this is just for today.
80K number might have no meaning (not close to any round number) in a year or two.

I so hope you are right :-)

[josegines]

and HODL til $80k -

why $80k?

right now BETI is about -2.09.
About 29K at BETI=0
80K at BETI=1.0 (as the last peak).

However, this is just for today.
80K number might have no meaning (not close to any round number) in a year or two.

Thanks!

but the curve is based on fixed parameters for several years? because in that case, would not it be logical to think that this curve might need a revision? How old are the parameters used to calculate the curve?

we may reach 80k, but it could happen with BETI = 2 in a new curve

[DieJohnny]

It seems more likely that bitcoin is on a logarithmic regression curve, like so:

For me the BETI deviation is a broken indicator. I like yours much better at this stage in Bitcoin's life. Breaking $4500 was the point that new models need to be considered.

[thepkbadger]

chart update, if anyone is interested...

[LUCKMCFLY]

It is interesting to see the different points of view that the technical analysis can give us, although as I see the market today, I think it is best to wait, because we are in a phase of accumulation, and the movements that occur from bitcoin upwards, can quickly fall because everyone wants to take profit ..

[birr]

The trend line looks too high.

[Cassius]

The range is moving down via a SQRT curve of sorts - we aren't ever breaking zero again and if we touch -.5, sell everything! I like this indicator, but it should somehow evolve away from log and towards a more gradual curve. I patiently waited for another 1.5 on the last run up without realizing this.

No need to abandon the log scale.  Only needs judicious choice of curve.  SQRT it is.
I ran some numbers on the price history and fitted a square root baseline.  Vertical axis, log of dollars.



This chart shows the relation between the price and the baseline over time.

This is a great piece of work and just what I'm looking for.
I'm putting together a series of indicators and trends for a news site, www.bitcoinbulletin.org. The idea is to offer fundamental/long term price development and projects as well as short term indicators. The first category will include 200 WMA, log curve and 'expert' predictions for context; the second signals like NVT and Fear/Greed Index.
I've been particularly interested by Peterson's work on Facebook/Bitcoin network size, time and price in this paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3356098 This has proven an extremely good baseline for bitcoin's 'clean' price (i.e. non-bubble price).
@birr do you have the equation you used for that square root baseline? I've been struggling to come up with something meaningful myself.
Appreciate the work you've done on this and thanks in advance.

[birr]

@Cassius
parabola has an apex, so there's a time zero on this graph, before which price is undefined.
My latest equation has time zero as Sept 22, 2010.  That is the day that realistic price discovery began, according to this mathematical scheme.
The equation for baseline (in red on the image) is
=0.0926237*(D^0.5)-1.57846
where D is days since Sept 22, 2010

If you want to extract a square root line from historical price data, the way to do it is swap axes.  Make time a function of price, and do a quadratic regression on the data (dead easy).  Then solve the quadratic.