Something I've expected for years, but have never had evidence... until now.

Or, alternatively, that submitting buggy patches on purpose is not research.