this post was submitted on 07 Jul 2023
1 points (100.0% liked)
/kbin meta
38 readers
1 users here now
Magazine dedicated to discussions about the kbin itself. Provide feedback, ask questions, suggest improvements, and engage in conversations related to the platform organization, policies, features, and community dynamics. ---- * Roadmap 2023 * m/kbinDevlog * m/kbinDesign
founded 2 years ago
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Edit, my post here was wrong @Pamasich appears to be correct, having checked the "top" sort order of /all I have concluded that it is according to Boosts less Downvotes, upvotes do indeed seem to be ignored.
For the record and to maintain the coherence of the message chain below this one, my previous post:
This isn't true. Top and hot work according to [upvotes plus boostsx2 less downvotes] same as reputation. This is easily proved by sorting by top and noting sometimes they are not in order of most boosts.
Sorting was actually made to work this way long before reputation was changed to match.
If I look at the top posts on /all/:
Total1958 |Upvotes711 |Downvotes43 |Boosts645 | LinkTotal1408 |Upvotes608 |Downvotes20 |Boosts410 | LinkTotal2062 |Upvotes1274 |Downvotes16 |Boosts402 | LinkTotal2198 |Upvotes1568 |Downvotes2 |Boosts316 | LinkFormula for total I'm using is
boosts * 2 + upvotes - downvotes.Really doesn't look to me like that's what it uses.
Though, now that you mentioned it, I do notice that rarely the posts aren't in order of boosts. Maybe the sorting algorithm uses an entirely different formula? Boosts do seem to be mostly responsible though, judging by what I see on /all/.
I was mostly going by what Ernest posted on codeburg four weeks ago.
https://codeberg.org/Kbin/kbin-core/issues/80
I'll admit that other than noting by eye
I've not actually tried making a spreadsheet to check the algorithm (...yet!)
Edit: looking at /all perhaps for articles the number of comments is counted too somehow?
Edit2: If you formatted your table as a code snippet it would probably work since it would use a fixed-width font.