In principle, the model fits so well in the past because it has of course been fitted in such a way that the deviations from the previous data are minimal. This allows you to "prove" any supposed correlations, especially if you allow for axis scaling.
However, it is true that, by nature, networks tend to satisfy exponential laws. Whether the price then also shows long-term exponential growth has nothing to do with this in the first instance.
However, it is true that, by nature, networks tend to satisfy exponential laws. Whether the price then also shows long-term exponential growth has nothing to do with this in the first instance.
•
1515
•
@RealMichaelScott I agree. However, the axis scaling is limited to the fact that both axes are displayed logarithmically. And in this respect, I find it interesting that the price more or less follows a linear line.
I agree that you can't predict the future based on this.
Especially since disruptive technologies usually follow an S-curve adaptation :)
I agree that you can't predict the future based on this.
Especially since disruptive technologies usually follow an S-curve adaptation :)
•
11
•@stefan_21 I haven't looked in detail at exactly how the axes were scaled. If it is only the pure logarithmization, it is of course a stronger correlation than if the axes were scaled accordingly beforehand.
•
11
•@stefan_21 I have now looked again and the correlation of the Bitcoin price should be:
1.0117e-17*(days since genesis block)5.82
But that's exactly what I meant: the 1.0117e-17 and 5.82 are simply scaling factors on a logarithmic scale to get a particularly good fit. In this respect, it is as I wrote in the previous comment: a rather weak correlation if scaling factors are allowed.
1.0117e-17*(days since genesis block)5.82
But that's exactly what I meant: the 1.0117e-17 and 5.82 are simply scaling factors on a logarithmic scale to get a particularly good fit. In this respect, it is as I wrote in the previous comment: a rather weak correlation if scaling factors are allowed.
•
22
•@RealMichaelScott Merci. Good point, I hadn't noticed that. Then there will certainly have been a bit of playing around to get that 5.82 factor.
•
11
•@stefan_21 Yes, certainly. The factor in front of it is also nothing more than an adjustment factor that fits the compression of the y-axis (i.e. Bitcoin price) in such a way that the deviation is minimal.
The website also claims that, in principle, a linear regression was carried out on a log-log basis.
The website also claims that, in principle, a linear regression was carried out on a log-log basis.
•
22
•