Timeline for Should answer scores decay? Discussion on making answer rankings for older questions more dynamic
Current License: CC BY-SA 4.0
10 events
when toggle format | what | by | license | comment | |
---|---|---|---|---|---|
Sep 15, 2021 at 19:27 | comment | added | Resistance Is Futile | I have been thinking about it a lot lately, but no matter how I turn it I couldn't imagine anything that would not make even bigger mess. And every other metric is just as fallible. On technical sites adding versions and version sorting could alleviate some issues that arise with changes in technology, but this cannot be applied on other sites. For instance, I cannot see how old math answer needs to age away to give more room for new one. | |
Sep 15, 2021 at 18:31 | comment | added | ColleenV | Yeah, I think every vote should count equally. I wonder if there's some combination of age, views, interactions with an answer (comments, edits), a "controversy" measurement, and other factors that we might use to adjust the order answers are displayed in? Maybe we side-step the thorny issues around adjusting scores and add some information in addition to score that humans cans use to search for and sort answers. Right now it's hard to use score in a meaningful way because it is dependent on views and age. | |
Sep 15, 2021 at 18:07 | comment | added | Resistance Is Futile | I was always against blocking people to see vote breakdown, but I also don't think that downvotes should weight more than upvotes (unless answers have same score). | |
Sep 15, 2021 at 18:04 | comment | added | Resistance Is Futile | I wouldn't mind if answers with equal score would have additional something to determine better rank. If I am not mistaken, such answers position is determined randomly. For me that is good enough, too and I don't think there is enough such answers (on high quality questions) that changing rank would be significant improvement for those. | |
Sep 15, 2021 at 17:58 | comment | added | ColleenV | Currently votes => score => ranking. What if votes => score + (???) => ranking? "Sort by votes" is actually "Sort by score"... answers with zero votes and answers with an equal number of downvotes and upvotes are ranked the same, but are they really the same? The drive-by visitor from a search engine can't even see whether the score is controversial or if no-one voted on it. I think you're right to be skeptical of an automated process, but what if the automated process is just a nudge instead of a "fix"? | |
Sep 15, 2021 at 16:59 | comment | added | Resistance Is Futile | To make it more clear, I cannot imagine any specific implementation where automatic decaying (or ranking) would yield good results. We can only manually detect and deal with aging answers. One thing in good direction has been done - unpinning accepted answer, but for the rest, some sort of manual "flagging" and marking would be necessary. | |
Sep 15, 2021 at 16:56 | comment | added | Resistance Is Futile | I am not sure what you mean by "little to do with votes and more to do with the score" as score depends on number of votes? | |
Sep 15, 2021 at 14:33 | comment | added | ColleenV | You seem to be imagining a specific implementation of how the score would decay, and I concede that making votes decay explicitly is problematic. Is there some more creative way than aging away votes that we could make the ranking more dynamic? What factors could we look at to detect if an answer is aging well instead of poorly? | |
Sep 15, 2021 at 14:32 | comment | added | ColleenV | I tried to scrub the "time" aspect from the question--initially I was thinking more about activity than age. It's discouraging to see a one link-only answer with a high score and a newer comprehensive answer with a low positive score and know that that the old answer will always rank above the more useful answer. What if it wasn't time based? What if the decay had little to do with votes and more to do with the score for ranking? | |
Sep 15, 2021 at 14:22 | history | answered | Resistance Is Futile | CC BY-SA 4.0 |