You are viewing a single comment's thread from:

RE: Announcing the Forthcoming Open Sourcing of UA, Free Accounts Give-Away, Unregistered UA Score & Rank Checking on steem-ua.com

in #ua6 years ago

@steem-ua is not a manual curation program and does not directly consider the content of the posts to be curated. Instead it considers multiple factors such as the UA scores of followers, upvoters, and (I believe) commenters to decide how much each user should be rewarded per post.

If the "one picture posts" are of insufficient quality, the idea is that people will, in theory, unfollow that user and stop upvoting and engaging. As a result, that user's overall UA score will drop and so will their reward.

Whether or not that works as planned is the subject of the experiment. :)

Sort:  

problem with considering the upvoters is that by using smartsteem bidbot, there care huge numbers of big names selling votes there so its bound to look like you are super popular with the dolphins! It its not considering quality of post, I dont see the point?

Posted using Partiko Android

There are a number of valid concerns about several aspects of @steem-ua and I think it's great that people care enough about Steem to raise them.

I see @steem-ua as a fun and well-intentioned experiment in attempting to reward organic behaviour, which is why I delegate. At this point, their intentions matter more to me than the actual results, and those intentions appear benevolent. Meanwhile, at the very worst, the current result is benign. I feel comfortable delegating knowing that I (and all other delegators) can always undelegate immediately if that changes.

Well said!
To share with you my take on the same topic: indeed @steem-ua is one use case / application (out of many!) of using UA data, and indeed the objectives of @steem-ua are "getting in & keeping in" users on the entire Steem Blockchain. To create a certain mechanism that incentivizes and rewards "good behavior" from a community perspective. If you shitpost, the value of that shitpost is still "something" but not as much as it could have been if you really gave it your best shot to produce top quality content. The algo identifies - pretty accurately as well - how others interacting with the post value that post.

For example, somebody in this comment thread wrote he/she is of the opinion that @berniesanders has a too high UA rank. I don't agree to that, because the entire Steem network is taken into account for him to have that UA score / rank. Does @berniesanders deserve an upvote, being a @steem-ua delegator, when "shitposting" (as he calls it himself) a picture of a toothbrush and a lamp, and/or create a contest / challenge for others to publish similar pictures?I'm of the opinion he does deserve our upvote indeed. You can agree or disagree or think whatever you like about his actions, but it is a fact he is influential on Steem, and that is what UA measures! Yes he is influential, and therefore yes he deserves a high UA_Account score.

Now, the interesting thing - as far as I'm concerned - is this: if @berniesanders decides to publish a masterpiece of content and spend a ton of time on writing it, then I am quite sure our algo is able to identify that that post of his deserves a higher upvote than his other posts. Intuitively, that justifies the algo functions as intended.

Also, suppose some random, unknown, rather new, not influential account delegating to @steem-ua would publish a similar example post with a challenge to post a picture of a toothbrush and a lamp: would that similar post deserve the same upvote? No! Because close to zero people know about that account, their post will not receive much attention.

And this brings me to an important distinction to make: @steem-ua is indeed an experiment to value "content quality", about any topic in any language, but not by the content alone, but also - or even mostly - by who (based on each post's author, voters and commenters) publishes and interacts with it (!!!).

A one photo post with (close to) zero additional textual content can have value, but who's to decide, and by which objective criteria, if that piece of content is "quality" content, and/or has "more or less quality" than a 500 word article? @steem-ua lets the giant network of follower relationships and the witness stake and their followings decide.

Is that perfect? No. Can we distinguish top quality content, always, 100% of the time? No. But it does work reasonably well.

Steem-UA disregards bidbots. The top x% at least

Posted using Partiko Android

so, how come it votes on one pic shit posts ?

Posted using Partiko Android

Great answer as many have been commenting on this lately with negative connotations. While I am sure it will be evaluated, it will be great to eventually see the correlations in the scores with metrics on quality and the such. Maybe something @abh12345 can look at with his Engagement League?