Alternate voting system and Steemit Posts with more votes than viewssteemCreated with Sketch.

in #steemit7 years ago

I noticed right now that most of the posts on my feed, at least on the first few pages that I checked have allot more up votes than views. What exactly does this mean? I am fairly new to Steem but I would assume that people are up voting posts without even reading them. I understand that many of these up votes are probably from bots.

How should Steem take care of this problem, if it is a problem at all?

Here are a few examples:

This post has 63 views but 83 up votes. Assuming every viewer voted that is 31.7% more votes than views. However I assume not every viewer voted so like much more than a third of the voters never even bothered to read the blog post.


This post received more than 3 times more votes than views.


This post has more than double the votes to viewers.

I realize that if that change Steem to only accept votes when a post is read bots will simply open the post first. So I am unsure how one would go about resolving this.

I have been thinking of an alternate voting system for a paid blogging system like Steem and I don't mind is anyone including Steemit uses this idea. The idea is that each post has a 5 star rating instead of just an up vote. You can only rate a post once the post has been opened since it does not make sense to allow votes on headlines anyway even if bots can still open the post and just vote afterwards. Of course it's not just bots that up vote without opening posts. I tried it myself when I first discovered Steemit to see what I could get out of it.

You may be wondering how that changes things. The trick is that your vote is measured against everyone else's vote. So if you vote 5 star and everyone else votes 1 star you wont get any or much curation reward as the reward must be linked to how close your rating is to others. If you vote 4 star and everyone else votes 5 star you will get some but not as much as those that voted 5 star but more than the person that voted 0, 1, 2, 3 star. 0 Star will be like reporting a post.

Another example is the average rating for a post is 4 star so those that voted 4 star will get maximum curation those that voted 3 or 5 star will get some. 1 Star voters will get less and 0 star even less or nothing.

This will encourage people to read a post and rate the quality of a post. So even if you disagree with an opinion post you might still rate the quality high because you know just voting 0 or 1 star on a good quality post you disagree with will not get you much curation. You have to think about what others will think of the post.

Of course you might think well just create more accounts to change the average. However the weight that your rating carries must still be linked to your Steem Power as it currently does so more accounts with your Steem Power divided amongst them will no give any added benefit to overall weight of a particular star rating.

One issue is that botter's will statistically analyse that maybe a 4 star rating will give them the most benefit and just always vote that. That can be countered by reducing payouts or voting power if you repeatdly vote the same star rating. It could also encourgage people to spend their voting power on lower rated posts to encourgae people to rate lower posts. A certain budget could be allocated to curation for every star rating too. So people who rate 1 star posts still get paid curation even though the post itself will not get a high payout as a result of the lower star rating.

Which brings me to that point. A posts payout will be related to it's star rating. So 1 star rated posts will receive say a fifth of the payout.

I could go on with the benefits to a voting system like this. And I know there are ways to exploit this too that must be ironed out. But ultimately I believe it will be a much better system. A good example is also the old Netflix star rating system. Where Netflix would display the rating it thinks you would rate the video. So I for see that this can also be applied to the blog posts. Show you what it thinks you will rate it so you will only see relevant posts. If you rate something 1 star it will learn from that and you feed will change to filter out similar content. This learning system could however be implemented in a future release. Initially I would not show the star rating until payout have occurred and maybe just show the vote count until then.

Please post your thoughts below. Am I missing something big? How could you fix that?

Sort:  

This would definitely be an interesting system to see implemented. However, with a large enough number of bots(so the majority of the up votes) up voting, in comparison to regular voters it would still cause the bots to win. Does this just make the system more complicated, and will it in the long run just cause an influx of bots in the system.

I think the point I am trying to make is a bot will not know what to vote and their vote will risk getting less i.e. wasted voting power is they are wrong so it would be better to read a post and vote correctly than just guessing.

but would it not be beneficial to just set the bots to automatically give it a 4 star rating. This would cause the majority of the count to be in the 4 star area unless it is really bad. Then it is just dialing in the algorithm as to which posts to automatically upvote. I see how this could potentially deter bots, but It really just makes the bot creators become smarter. Realistically there is no easy way to fully deter bots. My question would it really be worth it to go to such lengths to really try and stop botting

Well as long as voting is an inherent part of earning SP yourself. People will keep up-voting content they didn't even red.

You could remove the upvote button on overview page.

But people would just click the article and scroll down to up-vote, and still without reading.

If people really just want to up-vote articles without reading them there is no practical way of stopping that.

Hmm perhaps the addition of a timer could help.. And it would need to not be annoying, except for those people who scroll straight down and then upvote... The timer would take into account the number of words and assume you are a damn fast speed reader and work it out from there, only once the time has elapsed are you allowed to upvote and rate.