vectorbt
Vectorised backtesting library built on NumPy and Numba, designed for testing large parameter spaces fast rather than simulating one strategy carefully.
What it is good at
When the question is 'how does this behave across ten thousand parameter combinations', nothing else in Python is close on speed.
Scored against the five published criteria, identically for every entry, by the editors of this site. Scores are never sold and are not affected by any commercial relationship. Read the method.
What to watch
Speed comes from vectorisation, which makes path-dependent logic and realistic fill modelling harder. Know which question you are asking.
This section is not optional and is not negotiable. Every entry in the register carries one, whatever the commercial relationship.
Access it
No commercial relationship. vectorbt is listed on editorial merit and has no input into this description or its scores.