How does Nashpy relate to Gambit¶
Gambit is the state of the art software
library for Game Theory [McKelvey2016]. It also has a Python interface. It
handles \(N\geq2\) player games and is computationally efficient. It is a
much more mature piece of software than
It does however sometimes prove difficult to install (because of the required C libraries), in particular installation is not supported on Windows. In those instances you can use Game Theory Explorer which is a great web point and click Graphical User Interface (GUI) to Gambit.
The main mission statement of
Nashpy is to provide a
Python library that implements algorithms that are implemented using the
scientific Python stack (
This is motivated by the fact that I wanted a Python library (not a GUI as I am keen to teach reproducibly research methodologies) for teaching my Mathematics students. Using the Gambit Python interface is not sufficient for this as students need to be able to install it on their own machines (without difficulty).
All the algorithms in
Nashpy are implemented with readability as the
main motivation. This at times comes at an efficiency cost. For example,
support-enumeration builds the entire Polytope representation (using
scipy) which is not efficient.