To reproduce the results from the submission "Characterizing the Convergence of Game Dynamics via Potentialness", checkout the readme.md in matrix_game_learning/projects/hodge. You should find all relevant information there.

To run the scripts:
- install the package games_decomposition (instructions are also in the readme)
- run the scripts as descriped in in matrix_game_learning/projects/hodge/

The results of the "computation scripts" are stored in matrix_game_learning/projects/hodge/data/.
To generate the plots, you only have to run the "evaluation scripts".
Due to size constraints, the data on matrix games has to be generated first.