So I made some bold forecasts of the S&P 500 (in January) and the bitcoin exchange rate (in May) using the dynamic equilibrium model. The S&P 500 forecast is doing really well (second graph is a zoom in on the first):
In the bitcoin forecast, I made a bold claim about a turnaround in the price — and the turnaround came and was well-described by the model:
However it was followed by a shock (centered in early to mid August). This shock is of the typical size in the historical data (best seen in the logarithmic scale graph):
It's a bit clearer if we zoom in. The forecast worked remarkably well for the data up until the bitcoin fork on August 1st:
The most recent data looks exactly like a shock centered at 2017.6, or 10 days into August. It's width is about 6 days, so it's perfectly consistent with the August 1st bitcoin "fork". Here is the dynamic equilibrium model fit to the recent data:
There was no telling how big the fork shock would be using the dynamic equilibrium model, but you could understand the shape.
Update 5 September 2017
The Mathematica code for the bitcoin model is up on my Github.
Update 6 September 2017
Jupyter notebook/python IEtools code is now available on Github as well. Forecast does not include error in this version, however.
Update 5 September 2017
The Mathematica code for the bitcoin model is up on my Github.
Update 6 September 2017
Jupyter notebook/python IEtools code is now available on Github as well. Forecast does not include error in this version, however.
No comments:
Post a Comment
Comments are welcome. Please see the Moderation and comment policy.
Also, try to avoid the use of dollar signs as they interfere with my setup of mathjax. I left it set up that way because I think this is funny for an economics blog. You can use € or £ instead.
Note: Only a member of this blog may post a comment.