Hello, Karel & Community,
While the PTL database is amazing and pretty accurate (backtesting and actually trading the pair seem aligned in term of perf), I need to compute the cointegration p-value using Python for some reasons.
I've been doing a lot of research, but there are soooo many Python packages available out there to compute the p-value, that I would like to know which one you recommend, and why?
Examples of python packages providing the calculation of the p-value:
- One step only: statsmodels.tsa.stattools with coint()
- Two steps: statsmodels.api with OLS() + statsmodels.tsa.stattools with adfuller()
- scipy.stats with linregress()
- Etc.
