Hi Sergii, It turns the model of Trolololo is nothing else than a power law. He wrote the formula, log10(Price)=A*ln(days)-C but this can be rewritten as Price=10^(A*ln(days)-C) or Price=days^(Aln(10))/10^C or price=B*days^n where B=1/10^C and n=A*ln(10). It is a big deal because a power law first of all looks like a straight line in a log log graph and BTC looks exactly like that. .....
With all respect the formula in question is NOT a simple power law and its log-log plot does not look like a straight line
E.g. here is one recent version of such plot in log-log scale
Re: Logarithmic (non-linear) regression - Bitcoin estimated value
by
yurets_z
on 24/08/2022, 09:57:40 UTC
dragonvslinux Concerning the model, I believe the curve should be updated after adding new historical data to it. At least that is what I do to build the plot that I use for my forecasts.
Post
Topic
BoardSpeculation
Merits1 from 1 user
Re: Logarithmic (non-linear) regression - Bitcoin estimated value
Clearly it's not arbitrary when LookIntoBitcoin have modified the logarithmic curve, and haven't referenced the fact it's arbitrary or lacks any scientific basis. Probably because it's a mathematical based curve, nothing to do with science. The irony being that a logarithmic curve in itself isn't arbitrary, it's simply how you style it (connect the dots), even if the original remains arbitrary and unchanged.
You didn't get me. The Log curve is an approximation of price history (logarithmic regression). That is the scientific part. But the rainbow is made by duplicating that curve parallel to itself with certain intervals that are not somehow "scientifically" linked to volatility, ATHs ant Lows.
Post
Topic
BoardSpeculation
Merits2 from 1 user
Re: Logarithmic (non-linear) regression - Bitcoin estimated value
I reference this as the LookIntoBitcoin version looks more relevant, as factors in the March 2020 low more accurately, as seems to have been adjusted to include it.
You should always remember what is written on one of the referenced web pages: "The color bands follow a logarithmic regression (introduced by Bitcointalk User trolololo in 2014), but are otherwise completely arbitrary and without any scientific basis."
Post
Topic
BoardSpeculation
Re: Logarithmic (non-linear) regression - Bitcoin estimated value
by
yurets_z
on 24/05/2022, 14:14:41 UTC
Now it's OK. But maybe it's time to slightly update the model behind it.
Post
Topic
BoardSpeculation
Re: Logarithmic (non-linear) regression - Bitcoin estimated value
Hey I was wondering if you could update the bitcoin regression equation? It looks like you said in 2021 it's updated regularly but I'm seeing that the last update was in 2017.
If you click on the link within hotpassion's post - two post above yours, then you will see that the update of the graph goes all the way to today's date.
The graph is updated (current BTC price), but not the model from logarithmic regression.
