I do agree however that measures of LZ for example do peak with randomness (though there is a really cool old paper showing that this might not be the case, if I find it I will link it here). Measures are only as good as their utility, so one should always be cautious with what we infer from them.