How does Voting Power really affect curation rewards on Blurt? Help me find out!

Please comment here! I'm going to vote myself down below 50% in an effort to learn exactly how VP affects vote size and curation rewards. Does keeping VP in the 90-100% range maximize total payouts and curation rewards? This is the question arising from my recent posts about proper timing of curation (voting). Yesterday, I gathered some data with the help of my wife to find the answer, and the preliminary results are surprising. Now, I need help from the community to take it further and reach conclusions that will benefit us all!

Background

A week ago, I posted The Timing of Curation, where I discussed how Voting Power (VP) influences the value of our votes:

"Without curation, nobody gets paid, and good content is not created or discovered. But there has to be a limit to how many votes each user can dish out - otherwise, someone could drain the limited reward pool and monopolize the earnings. That's why the blockchain limits how many votes each account can make.

You start with 100% voting power (VP). With each full vote made, your VP decreases by 2%. Your VP increases naturally (heals) at a constant rate of 20% per day. That means you can make 10 full votes every day without depleting your voting power. If you're at 100% power and leave someone a full vote, you'll go down to 98%. If you vote again immediately, that vote will only allocate 98% of the rewards the previous vote allocated. But if you wait 2.4 hours (2 hours and 24 minutes), your VP will be back at 100% again, and you'll allocate the full amount.

The lower your voting power, the less ability it has to allocate rewards from the pool. For example, if your votes normally give content 30 BLURT (currently 14 cents), and you deplete your VP down to 50%, your next vote will only allocate 15 BLURT (currently 7 cents). As you can imagine, keeping your VP as close to 100% is vital, as that is what determines how much you (the curator) and the person who made the post/comment (the creator) are paid.

When your VP has healed to 100% for more than a moment, you are effectively wasting it. The platform as a whole doesn't care, because the reward pool - and the rewards paid out of it - are still the same size. It just means that you aren't participating in allocation of those rewards. The users you would have voted on don't benefit from your votes, and you don't get any curation rewards. So leaving your voting power at maximum slows your growth.

The absolutely most important way to increase your personal slice of the Blurt reward pool is to keep your voting power as close to 99.99% as possible, without letting it sit at 100%.

With an understanding of how this all works, a bit of practice, and dedicating some time to curating properly, you'll be able to make almost all your votes in that "sweet spot" of 95 to 99.99%. When you do that, you're throwing away almost none of your potential income, and maximizing the growth of the users you feel are contributing the best content to the Blurt blockchain."

That post was well-received by the community, and did not stir up any controversy. Some users thanked me for explaining it.

Yesterday, I did a post called The top 20 curators on Blurt, which ranked all accounts by the amount of rewards they were bestowing on content (which is also the amount of curation rewards they are earning). The ranking took into consideration not just how much Blurt Power each curator has, but also how much Voting Power they are currently voting with. That's because someone at 30% VP is missing out on 70% of their vote and rewards, compared with someone carefully maintaining their VP just below 100%... or at least that's the current understanding.

A couple people contradicted what most of us believed was already a settled issue. User @leifasaur commented:

"I was also questioning this concept of vote timing being so important and keeping things close to 100% as a way of earning more."

I invited leifasaur to test it out himself:

"Experiment with it! Run your VP down to 0%, and see how much your votes are worth. You will be giving out 0 BLURT per vote, and earning 0 BLURT as well. At 50% (like mk currently) you are giving up half your vote. At 90% you are throwing away 10% of your votes. The alternative would mean that VP does not influence rewards, but you know that isn't true because when you give out several votes in a row, you notice each one is worth less than the one before it."

He did some investigation of his own, and came back with:

"Well @drutter is right if he can truly keep his VP at 100% and just vote every 2 hours down to 98% but this is pretty hard to do" and "from 100-80% this is going to be negligible. For someone with a fixed amount of time that they can vote, using their 24 hour allowance of votes means more earning than letting their VP sit at 100% unused."

It sounds like he came to the same conclusion the rest of us did. Nobody ever said it was easy to stay in the sweet spot near 99%! If you're going to be away from the blockchain for a long time (for work, sleep, family, etc) then going down to about 80 or 90% makes sense, as you don't want to be sitting at 100% and letting your votes go unused. I covered that in my Timing of Curation post.

Interestingly, the platform's #1 curator @mariuszkarowski disagrees:

"timing doesn't matter at all or it can give you just a couple of %"

"there is no chance that one makes more than other one on curation, no matter what is your Voting Power when you are voting"

"the only scenario in which you lose curation rewards is when your voting power is sitting at 100%."

So, does Voting Power matter?

When trying to maximize your curation rewards, and how much your vote is worth to content-creators you're trying to support, does it matter what your Voting Power is? Or is a vote at 50% VP worth about the same as a vote at 100%?

We can have a look at the FAQ (which calls it by the old name "Voting Mana") for clues:

"Voting mana is like an energy bar in a computer game that goes down a little bit every time you vote. You start out with 100% voting mana. Every time you vote, you will use a small amount of your voting mana. As you use more of your voting mana, your votes will carry less influence. A vote with 50% voting mana left will be worth 1/2 as much as a vote cast with 100% voting mana. Not to worry, the network recharges your voting mana by 20% every day."

"Every 100% vote you cast will use 2% of your remaining voting mana. Your voting mana will recharge linearly by 20% each day. You can vote more than 10 times per day, but each vote will be worth less, and it will take longer to reach full voting mana again."

That seems straightforward, but I decided to run an experiment to find out for sure.

The Experiment

My wife @MediKatie is normally busy taking care of our babies, but she posts about 1 post or comment on Blurt each day. I asked her to make 6 comments last night, all similar in size, and all on the same post, so I can conduct my investigation. Keeping as many variables the same is important, so the comments each have a couple sentences of text and one photo. They're all in the same place, and were all made at the same time by the same account.

Next, I waited until my Voting Power healed to exactly 100%, and began voting on her 6 comments. I gave each of them a full vote, and took a screenshot of the information provided before confirming the vote. It took about 2 minutes to do all the votes and screenshots (about 20 seconds between each).

The results were not as simple as I anticipated!

At 100% VP, my full vote was worth 52.325 BLURT. I confirmed the vote, then immediately went to the next comment.

As expected, my VP had been reduced to 98%. What's interesting is the next vote was going to be 51.92 BLURT, which is only a reduction of 0.774%, rather than 2%. I confirmed the vote.

Two things were noteworthy about this one. One, my VP was only reduced by 1.95%, from 98% to 96.05%. Perhaps VP isn't reduced by an absolute 2% per vote, but by 2% of the current VP? I've never heard anyone talking about that before. (The experiment is still valid, it's just going to make the calculations a big more complicated. More on this below.) Two, the vote was going to be 50.863 BLURT, a reduction of very close to 2%, back in line with expectations. I confirmed the vote.

Again, VP had been reduced by 2% of the current VP, not a nominal 2%, so it went from 96.05% to 94.13%. The value of the vote was 49.828 BLURT, down about 2% as expected. I confirmed it.

Another loss of 2% of the current VP, and about 2% of the vote value. I confirmed.

The pattern is clear, with another 2% loss of current VP, and vote value. I confirmed this final vote.

Without confirming, I checked what my next vote would be:

Sure enough, VP had dropped by 2% of current Voting Power, and the vote would be worth 2% less than the previous. The pattern has been established, and we now have some data to analyze.

So we were right that the lower your VP, the lower the value of your votes. But there's a factor we didn't anticipate...

Voting Power drops by 2% of the CURRENT level, not by an absolute (nominal) 2%!

This was unexpected. I've never heard anyone else say this, even though for a long time I've been working hard to understand all the intricacies of Voting Power, curation rewards, and related topics. It's actually a huge distinction, and has massive implications for how we vote on this blockchain!

Here are my observed VP values:

100.00%
98.00%
96.05%
94.13%
92.26%
90.42%
88.63%

98.00% was expected. That's a 2% drop.

But the next wasn't 96.00%, it was 96.05%!

When first analyzing the results, I assumed my VP had healed by 0.05% between votes. So I calculated exactly how quickly VP heals (naturally increases). From the FAQ, "your voting mana will recharge linearly by 20% each day". Linearly means on a straight line - at a consistent nominal rate. That makes it easy to calculate.

20% / day
= 0.83% / hour
= 0.014% / minute
= 0.005% / 20 seconds

My votes were about 20 seconds apart, meaning my VP should have been healing about 0.005% per vote. So that doesn't explain why my VP only went down from 98.00% to 96.05%!

The led me to investigate if VP reduces by 2% of the current level, instead of the fixed 2% that everybody assumes.

98.00 - 2% (1.96) = 96.04

Holy shit! That fits the observed result of 96.05% almost perfectly. The other 0.01% was likely due to VP healing.

Okay, so what's 98% of 96.05?

---> 94.13

BINGO! That's exactly what was observed! And what's left after a 2% loss from there?

---> 92.25

The observed result was 92.26%, meaning my VP had healed another 0.01%.

At that point, I went and timed how long it actually took my VP to heal by 0.01%, and it was 43 seconds.

In other words, since my votes (and screenshots) took about 20 seconds each, my VP should have been healing at a rate of 0.01% every two votes. Exactly what was observed!

What's 98% of 92.26?

---> 90.41

My VP must have healed another 0.01 because the observed value was 90.42%. What's 98% of that?

---> 88.61

It had taken me a minute or two to navigate to an un-voted comment for the final reading of 88.63%, explaining the slight difference (caused by healing) of 0.02%.

So there we have it, confirmation that VP does NOT drop at a constant rate of 2% for each full vote, but by a very similar (but not identical) rate of 2% of the current level.

A careful reading of the FAQ actually hints at this, although it doesn't make it very clear, explaining why nobody has fully clued in before: "Every 100% vote you cast will use 2% of your remaining voting mana." The key word is remaining.

What are the implications?

It may seem trivial, but it definitely isn't.

VP heals at a nominal (linear) rate of 20% per day (which we've already determined is 0.83% per hour, and 0.014% per minute). Most of us, assuming full votes bring VP down by exactly 2% each time, have assumed that means we can dish out 10 of them per day. Indeed, many Blurt regulars have a daily routine of logging in, doing 10 votes to drop themselves down to 80% VP, and logging out. The next day, they've healed to about 100% again, ready for another round. While not a horrible tactic, considering it ensures you aren't sitting at 100% squandering your votes and curation rewards, it may not be the only way to do things.

Consider that at about 80% VP, full votes only "cost" 1.6% each, not 2%. Somebody hovering around 80% Voting Power still heals at 20% per day, but can actually give out 12.5 full votes (ie: 12 full votes and one vote at half strength) without any overall depletion of their VP!

And by extension, someone at around 50% VP only loses 1% for each full vote given out, meaning they can vote 20 times per day without depleting themselves. In theory, anyway, since I have never ventured down to those low levels.

That's where you come in.

The next experiment

I'm going to vote myself down to about 50%, taking data all the way. I want to see if full votes really only cost 1% each at that level, instead of the 2% when Voting Power is full.

Please, leave a comment (or three) below, for me to vote on. No links, no images, no videos, just a few lines of text. It can be on the current topic, or even something unrelated. I want them all here, on this post, so they're in the same place (to reduce confounding variables). Once I've got at least 30 comments, I'm going to start the experiment, voting down to levels I previously thought to be inefficient. I'll get screenshots, and record observations, and then post my results.

Hypothesis: At about 50% VP, considering 20% linear healing, and a 2% drop from the current value with each full vote, I will be able to give out TWENTY votes at 100% strength every day - instead of the usual ten - without overall depletion of my Voting Power.

Granted, it appears that my votes will be worth about half what they are when my VP is near 100%, meaning around 25 BLURT each instead of the usual 50. But if I'm able to give out twice as many, that means I'll still be handing out the same amount of payouts, and earning the same amount of curation rewards.

Downfalls? If this works, it means at 50% VP you have to find twice as many posts or comments to vote on, to use your VP as it heals. That might not be convenient for some people. Also, with half the weight behind your votes, you're not able to reward excellent posts the way you can at full power.

Advantages? If this is true, you don't have to worry much about your VP! It means that when you find a bunch of great content, vote away, and don't worry about going down into the 80s, 70s, or even lower. Your votes are worth less, yes, but you get more of them! It could even mean you can vote yourself down to 10% VP, allowing yourself to make ONE HUNDRED full votes every day, without any net loss to your Voting Power! Your votes would be nearly worthless, but all put together, they'd be worth just as much as they do when you're fully charged, and you'd be earning the same amount of curation rewards. Also, this ensures you're never sitting at the full 100%, effectively wasting your votes.

This would explain why mariuszkarowski feels he's not missing out on any curation rewards despite dipping down to 52% VP recently. It is dramatically different from how most people believe things work on Blurt, but after my initial experiment last night, it really appears to be the case.

In theory, anyway. Let's find out once and for all!

Thanks for getting through all that! I look forward to the outcome. It looks like I was wrong about part of my post on curation, so it will be nice to clear that up. After all, we must follow the evidence wherever it leads. Thank you for your help in the comment section below!

DRutter