9 ms·
Yes. Comment scores, user points and average per comment. The last was the most useful.
by jasonkester 10y ago
Yes. Comment scores, user points and average per comment.
The last was the most useful.
- danso 10y agoFWIW, the reasoning for removing the average karma per comment metric, which was about 2 years ago: https://news.ycombinator.com/item?id=8264220 https://news.ycombinator.com/item?id=8264220 > It's based on comment scores only. My sense is that it's an irrelevant distraction and we should get rid of it. For example, we caught a bunch of users who were gaming it by deleting any comments that brought their average down. We fixed that by treating deleted comments as comments of score 0 for average-computing purposes. It's an example of how, once you publish a metric, people start to care about it and do things based on it, regardless of how meaningful it is. Comment average used to be used by a few algorithms (like comment ranking) but we turned that off as an experiment a while ago and nothing seemed to get worse. If anything, I think it may have helped a little.
- inopinatus 10y agoThe interesting "missing metric" for me is comment heat. Rationale: I've occasionally made a polarising remark that appears to attract almost as many downvotes as upvotes. These remarks end up with maybe +3 karma after 24h of jiggling vigorously up & down but this doesn't represent the total number of votes. With any aggregate measure, the distribution of values summarized may be as revealing as the value itself.