Twitter thread https://threadreaderapp.com/thread/1385458194454097920.html
More than two years after this model was published, the price continues to track the model’s predictions with astonishing precision.
The price of bitcoin can be anywhere between $0 and $100m. That it continues to remain within the one standard deviation band of the model’s predictions for more than 2 years is about as likely to happen randomly as someone correctly guessing a private key.
The one standard deviation band is around $40,000 wide. The odds of the price being there randomly on any given day is ~40k/100m, or 0.04%.
The odds of it being in that band randomly for 800 days in a row is 0.0004 ^ 800 = 4.4 E-2719.
That’s 0. (2719 zeros) 4.
It’s very easy (and meaningless) to say “correlation is not causation”.
It’s very difficult to explain how this statistical impossibility keeps holding, even as it becomes less likely with every passing day. It’s even more difficult to make a model with a better track record.