pull down to refresh
97 sats \ 5 replies \ @Undisciplined 17 Mar \ on: The Math Behind SN's Web of Trust meta
Neat, so I got a trust boost within ~econ?
I talked to @k00b about allowing the components of v to be negative. Do you think that would help address anything? IIRC, I brought it up in the context of outlawing content from stackers with negative trust scores.
Good luck zapping anything in ~econ before me! Btw, I zap everything that's even remotely thoughtful. For things I find interesting, I zap more and for things I find interesting that took a decent amount of work, I zap a lot.
Yep, you get a trust boost in ~econ, and your zaps are relatively more influential in building other peoples' trust scores within ~econ.
Based on my understanding of the current trust graph, I think trust scores can be negative, but this would be rare because I think downzaps are pretty rare. I don't know if anyone has a negative trust score, but I think it's theoretically possible.
Currently, I wouldn't recommend outlawing people with negative trust scores yet, because I think there are a few technical issues to be worked out first. (There's a little bit of jankiness with the current trust graph calculations--I think they're directionally correct but the scales are hard to interpret and there are a couple of issues that I brought up in the github)
reply
I know k00b told me that trust can't be negative. I'm also pretty sure he said that trust scores only ever increase.
Maybe the outlawing thing doesn't make sense, since that's supposed to be for people who post stuff that tends to get downzapped, rather than people who zap antagonistically to the rest of us.
reply
Ah, I think that's right. Trust can't get negative because of a post-processing step in which the final trust scores are normalized between 0 and 1. I think the raw, unnormalized trust scores could get negative though.
I don't know about trust scores only ever increasing though. I don't think I see that from the code.
If you're interested in some of these details, here's a link to the github discussion that I started about this.
reply
Cool, I'll check it out.
reply
Interesting discussion. Most of those points seem pretty easy to fix. I like the idea of grounding it all in a behavioral model.
The point about long histories is interesting to me. I think your suggestion about only including items that both stackers are probably aware of would help a lot. Otherwise, my trust with new users (and theirs with me) will be weird.
reply