You are viewing a single comment's thread from:

RE: Steem Velocity Hardfork - Hardfork 20

in #steem6 years ago (edited)

Agree with you on the time window. The original suggestion from @blocktrades was to reduce it to five minutes, which I personally would prefer (if not shorter). However, this is a step in the right direction which addresses some concerns about making too extreme of a change and causing new problems.

I also think the original goal of addressing bots and autovoting is less needed now because there is so much more content. Originally it was quite possible for bots to instantly vote for all of the (five?) very popular posters and grab all the rewards (with superlinear rewards concentrating things even further). Now that is no longer the case, rewards are much more spread out and with limited vote power, deciding what to vote for is relatively more important then getting in instant votes.

I hope once we evaluate the effects of changing it from 30 minutes to 15 minutes we can consider further reduction.

Thanks for the feedback.

Sort:  

I am happy to see the reduction in the time window, however I would hope we could move towards the removal of it eventually. It does seem like it is a solution to a problem we no longer have(too many curators/bots, not enough content) and probably just over complicates the user experience now. I think it is important to not underestimate the negative effects of a mechanism which complicates the user experience and potentially adds a barrier for profitable curation. Personally I wonder if we could return to the no reverse auction , 50/50 split for rewards. It is much easier to explain and comprehend and I think it would help to reincentivize good curation over for profit self voting.

I see some value in a very short window. With no window you can have a rush of bots all flooding to get in the very first block of certain known-to-be-popular content, which could be viewed as a sort of unintended denial of service attack. A ramp that forces bots to pick what each believes to be the optimal time over a short window rather than all trying to be first at the exact same moment may be better for purely practical reasons of spreading out blockchain load.

But IMO this should all happen over bot timescales (something like 15 to 60 seconds at most) rather than human timescales, and ideally can be ignored by human curators rather than introducing the whole 'optimal time to vote' complication as is the case now.

Removing it entirely as you suggest may also be okay though.

I was thinking about suggesting the removal of the voting time as well, but like you said, if you remove that, the bots will have a huge advantage over the human curators. I've had GINAbot tell me that someone voted on my new post before it even told me that I had created one. The goal should be to leave a small period of time to let the humans get some curation, then open it up for regular votes and whatever else.

Thanks for sharing your thoughts. You've been around for a while and seen a lot of the changes, so you have a unique perspective on how to make it work well.

I wasn't around for the superlinear rewards, but it sounds like it was a whole different ballgame.

with limited vote power deciding what to vote for is relatively more important then getting in instant votes.

Also there are a lot fewer votes that we get per day this is even more important than "back in the day."

There probably won't be a system that is entirely fair. And there probably won't be a way to keep those who want to abuse the system from doing so, but we can try to make it a good overall experience for most people and we'll be doing fine. We're still making more money posting here than they are over on facebook. :)