So I was talking to my good friend @jrcornel today and we had a little discussion about downvotes. It started in response to @themarkymark's Proposal to return the 12 hour voting window gate.
I'm not sure why we got rid of the window and I don't really care. Downvoting has never been my thing. I'm not a fan of negative reinforcement and I've never seen content worthy of a downvote. To be fair I'm also not looking for counterproductive posts. I rarely ever even look at trending. Since downvotes became free under the EIP I haven't downvoted a single post (until today).
I just went down my list and started downvoting everyone in my feed just to test it out. You see, @jrcornel told me something that I didn't believe: that when you run out of downvote power it just starts using your upvote power, meaning we have more than enough power to downvote everything on the platform to zero; effectively every account has a 125% downvote pool.
It was my understanding that the downvote pool was made completely separate and 25% as large as the upvote pool. I had no idea how it even worked. Were downvotes 25% as strong as upvotes, or did downvotes simply consume 4 times as much voting power? After running these tests I got the easy answers I was looking for.
Logistics
So every downvote you cast pretty much takes away 8% of your downvote pool. This gives you 12.5 free downvotes. Once your run out of power it starts taking away your upvote power just like it used to before the EIP. It's important to note that those first 12.5 downvotes are full power, while that power starts to taper off when you dip into your upvote pool.
EIP
I've stated this multiple times, but I was vehemently against the EIP. Everything about it was potentially a bad idea, and to do it all at once was EXTREMELY risky. So far, it's turned out much better than I thought it would.
Truth be told, I thought we'd be seeing people selling their downvotes by now. They are just a free resource sitting around being wasted by many (myself included). Of course selling downvotes is a bit different than selling upvotes. If you sell your downvote you might make enemies with the person that gets downvoted, and then you'd make yourself a target for free downvotes.
The system is backwards.
The main reason I'm against the EIP is that it caters to the idea that Hive is only a blogging/social-media platform. We are manipulating our consensus backend to cater to the blogging frontends, which is completely backwards. How the frontends decide to interpret the blockchain (UX) should have nothing to do with our consensus.
We changed curation to 50/50 and added the downvote pool to "fix" the trending tab. Here's a novel concept: how about we fix the trending tabs of the frontends without manipulating our backend? No one is forcing high-payout posts to appear at the top of trending, yet the developers in charge of this network act like that is an inescapable fact that the entire network must be built around. It's a ludicrous mindset.
Imagine a prominent member of the Bitcoin community coming forward telling everyone Bitcoin Core code needs to be changed to accommodate the block explorers.
They would get laughed out of the room.
You don't change the backend to accommodate the frontend.
Isn't that obvious?
Yet that is pretty much exactly what happened with the EIP;
Except everyone took it seriously.
Eventually the entire system we've built here with the EIP is going to get gamed into the dirt. Downvotes will be sold to the highest bidder. Curation will get gamed with bots that cast upvotes on posts that have a high probability of getting upvoted (already happening). The convergent curve has no affect on the high-stakes accounts that we actually have to worry about. Money from the Proposal system will flow into the hands of people that aren't doing enough work to justify it.
Of course, I hope I'm wrong, but I doubt it. Regardless, I think Hive is still lightyears ahead of the "competition" so I'm not really worried about any of it. We'll figure it out, even if it takes a while. On top of that, every year that goes by the network gets more decentralized in distribution, so the system is far from totally broken. Baby-steps.
I guess the main point of this post was simply to point out how complicated this can all be. I've been here for 2 and a half years and I'm learning programming and I still didn't even realize the mechanics of the downvote pool because I didn't care to. I think it's stuff like this that really highlights now the unchanging nature of Bitcoin as the anchor of crypto is truly a feature and not a bug. Sometimes change is bad. If it's not broke don't fix it.
Conclusion
I've still never cast a single downvote on Hive. I did that all the testing on Steem because they still operate under the same rules. Not to worry, I reversed all the downvotes with upvotes immediately after logging the results.
Should we side with @themarkymark and stop people from upvoting during the last 12 hours before payout? Why not? Like I said, it's not my thing. It's certainly not going to interfere with anything I do here. If he wants to be the one to run a blacklist and return rewards back to the pool so be it. To each his own; every strong community is bound to have a full-spectrum of active players.
Is having a 125% downvote pool bad? No idea. In my opinion it's turned out so far to be the best thing about the EIP next to the Proposal system. Time will tell if these systems end up becoming corrupted or not.
It becomes clear to me that at some point I'm going to have to put up or shut up. If I keep making claims that the system is broken I'll eventually have to do something about it rather than point out the "obvious". No worries, more than enough room for White-Hats around here.
The EIP was never about "fixing" the trending page. It was about changing the game theory mechanics of the reward pool distribution to counter:
The change in the reward curve counters "hidden" low value spam posts and comments by reducing the ROI.
The 50/50 makes it economically agnostic to earn via posting or curating. Since voting requires less effort than posting it essentially incentivizes the former as opposed to the latter. This reduces the incentive to self vote 10 posts per day.
The downvote pool eliminates some of the opportunity costs of downvoting so it makes it easier for stake holders to remove rewards from less "honest" forms of posting.
Of course a side effect was that some of the more obvious "bad actors" were less likely to leech from the system and not show up on the trending pages but that was not the main point of it.
Posted Using LeoFinance
I'm fully aware of the merits of the EIP.
The problem is that even the definition of curation is "more visibility".
The foundational principal of curation is that higher payout = more visibility.
Meanwhile, things like reblogging aren't even monetized with curation.
I repeat, reblogging is the definition of curation but you can't get paid for it.
It's a backwards system.
Visibility is achieved by whatever merit the frontend decides.
It makes zero sense for the network to force curation on the backend.
All this being said there's really no point in arguing about it.
What we have is what we have and it's not bad,
I just think a lot of these mechanics are going to make sense in 5 years.
The front ends could probably do a better job at giving visibility to content based on different metrics. For the time being the current reward pool model is "good enough" as a coin distribution mechanism. We could do better. If the coin had a wider distribution the curation built into the system would be a close indicator of the popularity of content but it is waht it is I guess.
How difficult could it be to come up with a front end that you can customize to show posts by shares, comments, views and payout? Why do we keep using the same model?
Posted Using LeoFinance
I was opposed to the Economic Improvement Proposal, but I have to admit it's working incredibly well. Always remember: it's a miracle that the system is functioning at all, with so much freedom and quite a bit of money on the table.
@apofis is right that the EIP was about the actual distribution of rewards, not just the front page. You're right that front-ends should fix whatever's wrong with the Trending page. I'd propose showing posts upvoted by curation projects to people who are not logged in. Today, it's just a matter of aesthetics indeed. I don't care that high rewards go to meta posts about Hive and right-wing rants. It's not the first thing I'd want to show newbies, but it's not spam either.
Selling downvotes won't happen because it's easier to lease delegated Hive Power. And then the owner's name isn't on it. Leased HP is already being used for downvotes by a certain Chinese news account, and technically by me too, following @steevc's spamfighting.
I'll admit I don't understand all the ins and outs of the algorithm, but I work with what we have. If smarter people can find ways to avoid the bad guys abusing things then that's great. I am not finding so much to downvote on Hive, but that could easily change and we have to be ready.
Yeah, it is working quite well for now I agree.
Also, there are so many more important things to be doing rather than bicker about what we are doing with 8% inflation a year. 8% is nothing when we consider a killer dapp would send us flying higher than 10x
Free downvote pool is the unfortunate side effect of people unwilling to use their stake to help against actions similar to Haejin, bid bot abusers, etc.
Most people didn’t want to downvote before then because it earned them nothing. I was one of the minority that flagged things before that.
Now, it’s possible to push down undesirable content whilst upvoting those who need support.
It’s not perfect, but the only things we can change are incentives. The 12-hour thing has its drawbacks while address some other issues.
does it?
So far the free downvote pool has been nice to have around, but I'm not going to assume this will be the case 5 years down the road.
Yeah, if they make it that you can't upvote something 12 hours prior to payout, you'd be sitting duck to retaliation. At the same time, if you are engaged in downvotes, that should be expected.
And you'd be correct to assume things might not be as rosy 5 years down the road.
I'm glad to see even people i thought knew everything are wondering about this stuff!
Since 2016 i've been squawking downvoting sucks, like an angry parrot, and have never done a downvote. I think half the cost of a downvote should come back to the fucker who casts it, like upvotes.
I mean that's pretty much exactly what happened before they gave everyone a free downvote pool. Actually it was even worse, 100% of the downvote hurt you because you lost a 100% upvote.
sounds good!
Oh, good idea @edicted,...
ok people!! I have downvotes for sale if anyone is looking, send .13 hive with memo of the url to downvote.
Sidenote: I also don't downvote content... It's not my thing to punish others for whatever their effort may be... I won't reward it, but I won't police it either.
#DefundTheDownvote
lol defund the downvote police.
Good luck :D
Neat experiment....I'm glad I didn't post recently ;)
Relatively speaking upvote abuse is more burdensome for Hive than downvote abuse, even if being downvoted stings more.
Being downvoted for no good reason (because you got upvoted is a reason) deters individuals from staying on Hive and makes the atmosphere less friendly. However rampant upvote abuse (vote buying and trading and seeing crap content over rewarded) deters a lot of people from getting involved with Hive, too and makes it appear like a pyramid to no-Hivers.
For people who think it is unfair to give downvotes more time to shine, perhaps we can disallow downvotes for the first 12 hours after a post is made =p
Yeah, why not? It will help us get rid of a lot of spam with less work. Which will give everyone more time to think about other problems/solutions. This sounds like an "easy fix", at least for now. I'm not the one actually doing the downvoting but if this can help those that fight the abuse, then it should be implemented again... I think.
It's really about stopping those with sizable stake to be able to sneak in votes (even if at a loss) after day 6.5.
Example: someone with $10 vote getting zeroed at day 6.5 if they vote before.
Take a hit and vote at 6 hours before pay out. They get $5, but now it's harder and harder to zero them out.
But the problem then is retaliation can go unchecked near payout window.