Modeling Bitcoin Value with Vibes. This model has an R² of 0.97

This is clearly meant to be humorous etc, but just to explain a part of what’s happening here: Regressing price levels (integrated series) against each other does not lead to meaningful R^2 values and is considered a “chart crime”. See [1] for a detailed model.

[1] https://www.stat.rice.edu/~pe6/AOS1509.pdf

A nice article that can help people understand why time series regressions techniques requires special care and why R² is an outdated concept in statistics.

A 0.97 is good for a linear regression with 2 parameters. This model has 18 parameters, that is a lot, and perhaps more if you count the freedom to chose sin, cos, sqrt, and other stuff there.

It reminds me "discovery" of phosphine in Venus, where they were using a polynomial of degree 12.

This is hilarious. TL; DR 99% of pre-Bitcoin "quant" papers, and 99% of post-Bitcoin crypto financial talk.

Protip: you can fit any set of n points with a polynomial of order n-1.

This is why complex models are fraught with overfit concerns, and are poor predictors of future trends when any underlying assumptions change.

Don't tell this to the LLM crew.

TL;DR a parody of pseudo scientific discussion of the topic.