Tags: * poor ARR meta/review quality * improve matching * for tracks
In my experience from ARR reviews, the problem is really not the meta reviews, but the original reviews:
they are (surprisingly) bad, much worse than the average pre-ARR conference reviews. In particular,
they were very generic, written by people apparently not specifically concerned with the topic of the
paper. I believe that the Tracks idea - and/or suitable matching of reviewers with papers - is really
the crucial improvement that needs to be made.