I just read some content saying that using OLS on Bitcoin data is wrong and only quantile regressions are correct. Although I developed the Bitcoin Quantile Model, I don't believe this assertion is accurate. OLS is king and quantile reg, is way less used because it rests on more complex assumptions than OLS making it more prone to statistical failures. Further, the conditions where OLS fails generally also make quantile regression fail. What I see people do is overfit a quantile model by slicing and dicing the data and micro-optimizing the start day to get an artificially high fit, which shows a lack of basic understanding. I find it funny that people with no basic training feel confident enough to spam the feed with their statistical takes :)))(Sina 🗝️⚡ BI Report)