def test_boosted_exploiter_usually_reaches_destination(self):
"""Exploiter with boost=1 reaches destination >80% of the time."""
params = {'strategy': exploiter, 'q_boost': 1.0}
performances, agent_info = run_with_params(params)
pt = TrainEvalPerformance(None, performances, agent_info)
self.assertGreater(pt.prop_eval_reached_destination, 0.8)
def test_explorer_rarely_reaches_destination(self):
"""Random choices lead to destination <30% of the time."""
params = {'strategy': explorer}
performances, agent_info = run_with_params(params)
pt = TrainEvalPerformance(None, performances, agent_info)
self.assertLess(pt.prop_eval_reached_destination, 0.3)
