vonmises test in test_mixture_inference_quality.py fails

[riastradh-probcomp]

def test_vonmises_vonmises_model(self):
>       assert(check_one_feature_mixture(cycmext.p_CyclicComponentModel,
        show_plot=self.show_plot) > .1)
E       AssertionError: assert 0.093436305410546178 > 0.1
E        +  where 0.093436305410546178 = check_one_feature_mixture(<class 'crosscat.tests.component_model_extensions.CyclicComponentModel.p_CyclicComponentModel'>, show_plot=True)
E        +    where <class 'crosscat.tests.component_model_extensions.CyclicComponentModel.p_CyclicComponentModel'> = cycmext.p_CyclicComponentModel
E        +    and   True = <test_mixture_inference_quality.TestComponentModelQuality testMethod=test_vonmises_vonmises_model>.show_plot

[gregory-marton]

Where did the 0.1 come from? Is this flaky, or always wrong? What does its being wrong affect?