I agree with you. My anecdotal opinion is that commentary works well where the article is intended to impart information. A lot of HN submissions are of this ty

I agree with you. My anecdotal opinion is that commentary works well where the article is intended to impart information. A lot of HN submissions are of this type because they're posted precisely because someone learnt something from them and often lead to extra information being imparted in the comments. It's true, to a lesser extent in Guardian articles that are simply reporting. Where it doesn't work well, on HN or news sites, is where the article is intended to influence debate. Whether it's a monad swinging contest on HN or an analysis of the lighter side of Jeremy Corbyn, comments rarely seem to reduce the polarity. I think this is more of an issue for news sites simply because there is less reporting and more influencing done, somewhat ironically.