In Python, Gambit has a synthesised "outcome" class for strategic games derived from an extensive game, reflecting the fact that such outcomes are derived and not necessarily one-to-one with outcomes in the extensive game (because of chance moves).
This however is not implemented for (B)AGGs, and so accesses of outcomes on games with a (B)AGG representation result in a RuntimeError (for example in using .to_arrays).
In Python, Gambit has a synthesised "outcome" class for strategic games derived from an extensive game, reflecting the fact that such outcomes are derived and not necessarily one-to-one with outcomes in the extensive game (because of chance moves).
This however is not implemented for (B)AGGs, and so accesses of outcomes on games with a (B)AGG representation result in a
RuntimeError(for example in using.to_arrays).