pull down to refresh
463 sats \ 35 replies \ @ek 12 Jan freebie \ parent \ on: Stacker Saloon
Great write-up :)
However, regarding the trust scores: there isn't a single trust score per stacker. Trust scores are calculated between stackers.
If stacker A zaps what stacker B has zapped before, then B earned some trust from A (but not vice versa).
We do this to show everyone more of the content they like. If you use an inkognito tab, you should see that the items on the frontpage are ranked slightly different.
there isn't a single trust score per stacker.
really, but I think each stacker have their own trust score, but then in posts and comments ranking wise, then it's calculated among stackers? 👀
I found these when I was writing for my new post - are they outdated?
zap_trustis your trust which is only earned by historically voting for things other stackers deem are good.
(sum((1/(ln(zap_order+e))*log(zap_amount)*relative_item_favorability*(ceiling(zap_trust*3)+1))/sum(numerator across all stackers).
(ln(zap_order+e)) - this term change means you'll still be rewarded more for zapping early, but no longer disqualifies you for rewards if you aren't super early.log(zap_amount) - this term change means you'll still be rewarded more for zapping more, but no longer disqualifies you for rewards if you zap slightly less.(ceiling(zap_trust*3)+1) - this term change means you'll still be rewarded more when trusted, but no longer disqualifies you for rewards if you are low trust or untrusted.maybe @ekzyis could help us to decode it!
reply
really, but I think each stacker have their own trust score, but then in posts and comments ranking wise, then it's calculated among stackers?
Since gently personalized feeds, we use a matrix [0] so every stacker has a trust score from the PoV of another stacker. Before that, we all just saw content from the PoV of @k00b's account iirc.
I found these when I was writing for my new post - are they outdated?
However, I think it's good or at least very close [1]. The only "error" is maybe this:
zap_trustis your trust [from the PoV of another stacker] which is only earned by historically voting for thingsother stackers deem are goodthat stacker deems good.
Afaik, we didn't change the formula how content is ranked. Only how
zap_trust is generated.And I realized that what I wrote yesterday wasn't true:
I'm so bad when it comes to our own reward system. That's probably the part of the code base that I have the least experience with.
How you ranking works is the part of the code base that I have the least experience with. You always ask the best questions @Natalia! lol
When I have time, I'll look more into it today to give you an answer like here.
Also, thank you very much for writing these guides! They are very helpful, even for me who works at SN lol. SN is like an organism. It sometimes starts to do things even we, "the architects", didn't expect. And this is good and fun! SN lives from the community doing whatever it wants with SN. We just want to give you tools so you can do more awesome, weird, crazy stuff :)
And as you mentioned here:
Not gonna lie, it takes TIME to write good and original content!
It does take time. That's what I realized last time I wanted to write about ranking (but I guess I also took it way too seriously since I wanted to include the evolution of ranking, too).
And I think you are better at writing such guides than me. :)
[0] no, not that kind of matrix; this kind of matrix
[1] code is the source of truth and I didn't check before writing this reply since else I might have disappeared for an hour or more.
reply
Since gently personalized feeds, we use a matrix [0] so every stacker has a trust score from the PoV of another stacker. Before that, we all just saw content from the PoV of @k00b's account iirc.
oh, interesting 👀
If stacker A zaps what stacker B has zapped before, then B earned some trust from A (but not vice versa).
so if you zap the good stuff before others, you keep earning trust this way?
Also, thank you very much for writing these guides! They are very helpful, even for me who works at SN lol. SN is like an organism. It sometimes starts to do things even we, "the architects", didn't expect.
Team @sn keep building cool things, I enjoying learning how to play with it. 🚀
reply
reply
What do you see? :)
Maybe you can explain it to me and then I only have to correct a few things which is indeed a lot easier than me trying to explain everything lol
I call this the StackOverflow Effect
reply
I don't have time to watch anything :)
It was for Natalia to see, not me :)
reply
reply
Fair point!
@Natalia, how good is your SQL?
reply
bad, as in no experience at all. 🥸
How earnings (used to) work: 1/3: top 21% posts over last 36 hours, scored on a relative basis 1/3: top 21% comments over last 36 hours, scored on a relative basis 1/3: top upvoters of top posts/comments, scored on: - their trust - how much they tipped - how early they upvoted it - how the post/comment scored
Now: 100% of earnings go to either top 33% of comments/posts or top 33% of upvoters
about this, I feel like it's still calculate over last 36 hours? I remember like there was one time, a post of mine got rewarded two days in a row. 🤔
reply
Yeah, those are still relevant but you missed the latest release affecting ranking (iirc) from October 2023:
most of us will see the same content but it'll be ranked to suit your zapping preferences ever so slightly
But not blaming you, our search is really bad. I am even surprised that you found those old release posts, lol
The matrices are a very clever idea. This way we can have relative trust to other stackers. nice
That's very interesting.
If we implemented the trust score idea from these cowboy hat ideas, then we'd have different colored hats depending on who's looking at us.