zlacker

[return to "How Hacker News ranking really works: scoring, controversy, and penalties"]
1. swomba+ab[view] [source] 2013-11-26 11:45:02
>>jseip+(OP)
Hilarious that the original article was flagged off the front page, but this one isn't...

I find it very disheartening that the negative voices are being given so much weight. Everything that's worth doing will have detractors, and when it's something really worth doing it will have vocal detractors. Back when I had comments on my blog, every article I wrote that was any good had at least one person commenting that I was a moron or some equivalent statement.

Great things arouse passion - on both sides.

Giving 10x the power to the people on the negative side just creates an environment where new ideas are discouraged, where important but difficult discourse is pushed aside, where things of true import are penalised out of the group's attention by a few detractors.

There does need to be a system for flagging and removing spam articles, but if this system can (as it plainly regularly is) be co-opted to remove articles from sight just based on not liking them much, then it is broken. The people who have flagging powers are not responsible enough to use them wisely, perhaps.

I see at least one simple solution: lift the flagging privileges so it only becomes available to a much smaller segment of the population. Perhaps making the limit 10'000 instead of 500 would do that. That would still include hundreds of people, based on a quick extrapolation from https://news.ycombinator.com/leaders ). An even better model would be to make it dynamic - perhaps the top 200 commenters...

◧◩
2. ximeng+ek[view] [source] 2013-11-26 13:45:03
>>swomba+ab
Or open up the system more and allow people to add their own algorithms to weight how they want not how pg wants.
◧◩◪
3. obsess+an[view] [source] 2013-11-26 14:13:37
>>ximeng+ek
Yeah. I really would like to see a version of HN where there are no arbitrary penalties applied.
◧◩◪◨
4. eksith+Nn[view] [source] 2013-11-26 14:18:06
>>obsess+an
But we all already have that. Browse here : https://news.ycombinator.com/newest with show dead and viola! You have a raw feed with no penalties.
◧◩◪◨⬒
5. pbreit+lU[view] [source] 2013-11-26 18:57:30
>>eksith+Nn
That's not helpful. What night be nice is a URL with the voting and time factors but not the penalties.
◧◩◪◨⬒⬓
6. napole+oM1[view] [source] 2013-11-27 06:00:19
>>pbreit+lU
I just built this: http://news-yc.appspot.com/

It pulls https://www.hnsearch.com/bigrss and re-sorts it with the basic formula (but without penalties). If people like it/return to it I'll clean up the UI for better readability/mobile use (and display comments there instead of linking back to HN).

EDIT: Interestingly, without penalties it looks like this thread would no longer be on the front page.

◧◩◪◨⬒⬓⬔
7. napole+k42[view] [source] 2013-11-27 13:06:37
>>napole+oM1
FOLLOW-UP: My note in the edit was due to a mistake in the implementation, kindly pointed out by OP via email. It is now fixed, and this thread appears in roughly the same place on both sites (there is some lag between the RSS feed that I use as my source and HN proper, so I wouldn't expect an exact match even if I did no re-ranking at all). It is, however, easy and interesting to spot which articles have been penalized; there's an NSA article at the bottom of the front page right now that's at the top of my penalty-free page.

I'll be modifying my page to highlight penalized items and calculate their penalty when I have a bit of time.

◧◩◪◨⬒⬓⬔⧯
8. shoope+N62[view] [source] 2013-11-27 13:42:22
>>napole+k42
Thanks, very nice. How do you get story point scores from the RSS feed? The last time I looked at it, it was missing that crucial bit of information, and I don't see it now either. Or are you using parsing or another data source to get at them?
◧◩◪◨⬒⬓⬔⧯▣
9. napole+O92[view] [source] 2013-11-27 14:28:27
>>shoope+N62
Thanks! https://www.hnsearch.com/bigrss adds information to http://news.ycombinator.com/bigrss using the HNSearch API.
[go to top]