Hey, @forykw.
The stake weighted system for reputation ranking would work, if it weren't for human nature. And since I don't know how you overcome that, without some pretty motivating incentives, we're always going to see bought reputations, especially when people are less likely to pay bidbots to up their ranking as they are to do it increase rewards.
So, as long as people can upvote themselves, buy votes, or form a circle of votes, they can effect their own rankings far quicker than doing it the intended way will be.
However, I'm not convinced that the steem-ua way of measuring uers authority, which includes both a measuring of followers and gives weight to witnesses, is any better. The follower metric is broken in and of itself. People have followed to be followed, and are following so many others that the can't possibly see everything that comes through their feed, to the point that they ignore it.
And of course, a very small few in comparison to the total active user base are currently witnesses, so throwing that in not only ensures witnesses will be higher ranked, but it changes what the reputation rank was meant to be—a way to more easily find valuable and trusted content.
Which is the third metric. Create good content, get active people to follow, fire up a witness server, and you've got it made! Well, except, content is not king here, the upvote is, and so people aren't incentivized to curate as much as they are to find different ways to passively grow their stake. As long as that's the case, I don't think there's going to be a way to adequately measure value or trust.
Thanks for your view.
I used to think somehow the same "bright" way you are seeing things here, in the past, and then, year after year on the blockchain, I got faced with the reality that I would never see. This is a dream come true, but its not all about dreams coming perfect (or sometimes nearly far from it).
Without proper financial support, it's very hard to make things VERY right the first time, especially with a technology where the first purpose is to "experiment" and has the "embedded" idea of being constantly on "beta" stage.
My point is, I do agree that is important to state when things are not proper and what's not working or we think it does not work. Although, I feel the best juice of these experiments that can be taken into "consideration" is the things that do work! And this is the reason why sometimes I have this wish of "seeing more" about it or "accelerate" the progress of the experiment (if that does not change the path).
The current method of "reputation" is not even one to me. So, I am not considering it anymore for that purpose. On the ua side, what I like is the fact that we are "trying" something new. Does not mean straight away it will work, and I am also not saying it will work or not. Just that it is something different being attempted with the aim to address a reputation problem, and that for itself has some value on this platform (because is of interest of everyone).
Around this subject there is also other complexities around the equation... which are for example, questions like... "What should be reputation of a STEEM user?" or "What a user consider being highly reputable?". All very important to clarify otherwise nothing can't be really developed on top of the rationalism.
It's also important to think and realise that these things will always have to fight exploiting and gaming scenarios. But, it's also important to quantify how much of these we see... and in my view, that can also be seen as an improvement worth considering.
Finally, the reason I think the current weighted system does not work in my view is due the fact that when someone has high/low reputation, all is being considered is the accumulated voting for that user... and not the temporal relationship of that weighted value. Being considered "high reputation" forever for me is not a valuable credible input.
What I think it would work (my own view), is a system that has a depreciation factor along side with a dependency of what is "being considered" an high value reputation event. With the combination of these two, gaming and buying votes will also happen, but because the behaviour of people change and adapt, so the scoring of that representation would change. Making it very hard for static "exploitation". Yes, AI might be able to find a way, but before we reach that arena, I think the added variability, introduced with time would really make things very different.
What you think?
I'm afraid I don't know how to solve the issue, but I do know where I think I want things to go. Maybe knowing where we wan to be and working backwards can help. At any rate, the idea is to be able to look at the reputation of someone and determine whether or not they are a user who is adding value to the social media side of the platform through their posts, and I would add, comment and curation. There are other activities on the blockchain, but the reputation rating was supposed to be set aside for those who actually participate on the blockchain and can be considered a reliable source of information, value, etc. I think you need to include curation and commenting too, but again, I don't know how that happens.
Anything else, such as witnesses working on projects that are blockchain related, or community devs providing other tools or apps, but not specific to the post side of things, maybe they need their own recognition of some kind, but it would be nice for those who are putting in the time on chain to be distinguished from those who are not.
I think the upvote is over used. It should serve the function of curating, and that's it. I don't know what else to do, but having virtually everything tied to upvoting isn't the answer.
As it is, I would rather we scrapped it. I'm not into flagging that much, so not having a reputation would be fine, because then people could still function instead of being shunned into negative territory, and would only lose potential rewards. There needs to be a different way to deal with spam, and so far, the best way is having Steemit Inc remove their delegations. When they pop up again, take it away again.
Thanks. Valuable opinion. Will take it into my considerations. Cheers
Posted using Partiko iOS