You may have missed some details in your code about axelrod HOT 1 CLOSED

Thanks for the comment, always good to have someone check/question the accuracy of the code base.

On line 905 you'll see the application of a transformer to ensure the strategy defects on the last two plays.

There are a couple of reasons it is implemented like this (essentially knowing when the last two rounds happen is not always straightforward).

You can see in the tests specifically here: https://github.com/Axelrod-Python/Axelrod/blob/dev/axelrod/tests/strategies/test_axelrod_first.py#L664 that the strategy does act as intended.

        opponent = axl.Cooperator()
        actions = [(C, C), (C, C), (D, C), (D, C)]
        self.versus_test(opponent, expected_actions=actions)

        # Cooperator Test does noot defect if game length is unknown
        opponent = axl.Cooperator()
        actions = [(C, C), (C, C), (C, C), (C, C)]
        self.versus_test(
            opponent,
            expected_actions=actions,
            match_attributes={"length": float("inf")},
        )

The first versus_test ensures the last two interactions are (D,C), (D, C) so that the player in question defects the last two turns.

The second versus_test shows that when the length of a match is unknown the player in question just cooperates.

from axelrod.