Blurt Proposal: to Change the Vote Dust Threshold to Zero

I am impressed with the speed at which the Blurt team make decisions! However, I'm finding myself scrambling to write things up properly - and to triple-check the mathematics. This is an important change that will make the voting experience of the vast majority of Blurters far more meaningful.

I have opened an issue on GitLab: https://gitlab.com/blurt/blurt/-/issues/71

There is a parameter known as BLURT_VOTE_DUST_THRESHOLD that is currently set at 50 million rshares (defined here). I propose this should be set to zero. What it currently does is to decrease the value of everybody's vote, with far too many "dust votes" being created. The Blurt transaction fees now deter users from casting microscopic votes, so there is no need for this "vote dust threshold". Let me explain this in a bit more detail.

What does BLURT_VOTE_DUST_THRESHOLD do?

Let's go back to Steem for one moment. The STEEM_VOTE_DUST_THRESHOLD still exists on both Steem and Hive (as far as I can tell) and is also set at 50 million rshares.

Before HF20, this threshold set a minimum bar to how small an upvote could be cast. If you tried to cast a vote worth less than this threshold, you'd get an error message suggesting you power up more coins. At that time, this was equivalent to a 100% vote with just 1.2 STEEM of voting power. This is very small for most people and many users never even noticed the existence of such a threshold.

However, HF20 changed how this works; the parameter kept the same name but was used in a different way. From HF20 onwards everybody's vote was reduced by 50 million rshares. Again, few active users really noticed (beyond the HF20 RC glitch) but the new protocol allowed everybody to cast a vote, even if such a vote was worthless. Thus, instead of encouraging people to power up, the new system allowed users to waste their voting power and their mana - and clog up the chain's bandwidth.

To make a pole vaulting analogy, instead of setting the bar at 50 million rshares, HF20 made everybody's pole 50 million shorter! Some people ended up with poles the size of toothpicks - and were still allowed to participate.

HF21 did nothing to change this. Amid the confusion of what a convergent-linear curve actually looks like, very few people noticed that this vote-dust-threshold was never removed (or set to zero). This made the then new reward curve even more unrewarding for anybody at or below dolphin level. As an aside, I am helping Blurt overcome all such economic issues and will post about the reward and curation curves very soon.

So, this BLURT_VOTE_DUST_THRESHOLD is a legacy parameter from Steem days. On Blurt, it is having a pronounced effect on vote values because the formulas behind the scenes have been changed. Those changes, such as the removal of SBD, are designed to make Blurt easier to understand and leaner in terms of the calculations it needs to perform. However, there are consequences that sometimes only manifest when live.

Blurt could remove that parameter entirely, but setting it to zero will have the same effect. From what I can see, it is only used once in the calculation of rshares here. However, I shall leave such code-hunting to the coders! Just setting it to zero is so much less work.

What will be the consequences?

As I said above, the most obvious consequence is that those Blurters who have less than about 4,000 BLURT will see a significant rise in their vote values, and most "dust votes" should disappear. Everyone will actually see a slight increase in their vote values, especially those curation accounts who feel they need to spread smaller votes. This should also encourage users to power up more BP.

One consequence of increasing the vote values is that the Blurt reward pool will start to come down in size. Don't Panic! The Blurt rpool is currently at a very high value compared to Steem/Hive. Coming down is a positive sign of increasing valuable activity.

The reward pool mechanism works well and will eventually settle to a new stable band. We saw this after Steem HF19, when that reward pool started to shrink dramatically, with many wondering whether it would totally evaporate! The reward pool acts like a carburettor that adjusts the flow of coins into the economic engine. So long as coins are being minted, it can never drop to zero. I estimate it may take about 2 weeks for it to reach a new steady-state.

Between now and then I shall be working on the reward curve and the curation curve. There are choices to be made, such as whether to change curve entirely or change some of the parameters. The overall aim remains to make the Blurt economy simpler, yet still robust, flexible and more rewarding for more members.

If you have any questions or insights, please leave a comment.

---

_{images: erm... why?}

---