apsk32@apsk32
POWER CURVE + TIME DEVATION IS ALL YOU NEED
Business cycle models, M2 models, liquidity models, gold models, exponential decay models, quantile models.. they all go smoked in October. Analysts are overfitting. They project their models forward 10 years but have to change them every 6 months. "Most accurate," "best," blah blah blah. A model that needs a major revision twice a year is useless.
This baby here didn't need an update. It's robust. It doesn't over-fit. Simple. Elegant. 8 lines of code.
Import the price data, fit to a power function, measure time deviation, end.
Time deviation is physical. Is 600% over trend a lot? I don't know. Maybe? What was it last cycle? "Years ahead" tells you how long it could be before you see another ATH. Is 5 years ahead of the curve a lot? Yes. You could get a bachelors degree, you could meet a girl and get married, your 13-year-old will graduate high school in 5 years. IF YOU TRUST THE CURVE, then time deviation is all you need. Even if you don't know it, you measure risk against your lifespan. This natural tendency stabilizes the metric. "Years ahead" will remain a fairly constant distribution until the power trendline breaks.