3 ms·
Saying that a 20pt comment is better than a 5pt comment because solely of its intrinsic quality is confirmation bias. There are many factors which help determin
by minimaxir 10y ago
Saying that a 20pt comment is better than a 5pt comment because solely of its intrinsic quality is confirmation bias. There are many factors which help determine the score a comment receives, some of which are noted in the paper (time-of-day comment was made, time comment was made relative to the time original post was submitted, the score of the parent, etc.)
For example, this comment made today, which is literally just a quotation, is at 13pts currently: https://news.ycombinator.com/item?id=14178760 https://news.ycombinator.com/item?id=14178760
But it was a quotation in response to a top rated comment in a top rated thread.
- derefr 10y agoOh, sure, a factor of 1-10x is noise. But if you normally write comments that earn 1-5 points, and then one comment earns 300pts, there might be something to the way you wrote it. Or might not. It might just be that you wrote the rebuttal everyone wanted to the stupid comment in a very popular thread. But you should try to figure that out, because if—after eliminating sources of error—you're still left with an anomalously-high score, then the cause was probably you. More than just with highly-positive scores, though, I also tend to analyze the type of writing I put into comments that ended up with negative scores. More often than not, it seems that my own problem is that I didn't sprinkle in enough "hedging words" (like the 'more often than not' in this sentence) and then someone came along with a single counterexample and a downvote storm ensued. I wouldn't have figured out that people had that reaction if the points weren't there.
- GrinningFool 10y agoHedging words that show awareness of the difference between opinion and fact in particular seem [1] to be pretty key here. I think[2] analysing which comments get up voted can be very helpful if you're doing so when looking at writing style. but when approached with an eye towards content, it seems[3] like it would lead to the echo chamber effect that often gets raised as an issue. [1] there goes one of those hedging words now [2] oops, there's another. [3] and a third. It would be interesting to see the ratio of filler/hedging/weasel words to votes in HN vs elsewhere.
- jononor 10y agoA point score normalized according to the total number of points in the thread could at least make it possible to compensate for the 'general popularity' variation of different articles.