Also makes me think there might be some convention that should (not required) be applied to assertions that they should be phrased as null hypotheseses.