Calling Curating Curators

in #steem5 years ago (edited)

One more time... just because I keep seeing it come up... let's just talk about curation...

By the end of this post, you'll understand what is meant by the following:

  • Vote before the hump, but not too early.
  • You don't need to wait 15 minutes to validate someone with a vote.
  • Find an undiscovered gem, then court a whale (curation group).

Let's get started. First, let's talk about what portion of the curation rewards is yours. Ignoring the early penalty for now, it depends on the following three things:

  1. The total vote value before your vote.
  2. Your vote value.
  3. Total tote value at the end.

Without digging into formulas, I'll just give you the intuition behind it. Clearly, the more (2) and (3) are, the more you get.

Now, the more (1) is compared to (2), the less you get, and by quite a bit. If you want some rule of thumb, let's just say if you're curation-minded you'll want to place your vote so that (2) is at least half of (1). I've heard other fractions thrown around, but the point is that it needs to be comparable amounts.

And that also means that if you're looking at a post whose current value is way more than what your vote value is worth, just forget curation, and vote solely on the basis of whether you like the post or not. You are not getting a significant chunk of the total curation anyway. But remember that 75% of your vote value is going to go to the author.

(Edit: Thanks to @jga and @jacekw for pointing this out in the comments:

Good news is that at the worst case, if you vote late, even if there are a lot of votes before you, you will earn curation worth at least 1/8th of your own vote value. See this post for details.)

To summarize this section, it is simply this:

  • Vote before the others do, or Vote before the hump...

However...

Adding Back Early Penalty

If we simply left it at that, you can see an obvious problem: If you know an author is popular, then everyone is just going to want to be the first to vote on anything the author writes. This rewards bot behavior at the expense of rewarding real curation efforts. So they added a penalty.

If you vote a 0 minutes, you get none of the curation. If you vote at 15 minutes or later, there's no penalty, and you get the full amount of your curation portion. In between is linear, so that if you vote half-way between, you retain half of the curation.

Vote too early, and you give up a fraction of the curation you would have earned. And so now it becomes a psychology game, if we're talking about popular authors. How much curation will people be giving up to try to front run others? As you can see, there's no right answer, because it depends on everyone else's behavior.

I remember someone going "Vote 14 minutes! 14 minutes will make you bank!" However, this is fairly shallow, as it assumes everyone else has the mantra "I need to vote at 15 minutes so that I don't get penalized on curation". No. It depends on what other voters you expect to show up, and you need to keep in mind the previous section. Additionally, if you have a larger vote, you can afford to wait later in the interval, allowing smaller votes to pop in, and still reap a significant amount of curation. In some sense, the early voting penalty allows smaller voters to get more of the curation than they would otherwise get without the penalty.

Anyway...

  • Vote before the hump, but not too early.

"Back to the Pool"

You might have heard this phrase, and you also may have noted that you get penalized for voting early. So then you might wonder, if you get penalized, does it go anywhere? Well, it just gets returned back to the reward pool to be used in the next block. So don't worry, the money doesn't disappear or get burnt or anything like that.

If you put together all of this information, then there is one very simple consequence for the author:

  • The author gets 75% of the value of incoming votes, no matter when the votes are cast.

(If there's no self voting. Otherwise, treat the self-vote from the perspective of a curator if you want to know maximization strategies)

So you see? If you want to support an author, just throw your vote in. It simply does not matter. Yes, the presence of early votes on a post may deter future voters that have purely curation in mind (in favor of voting others that don't have early votes I suppose), or possibly those using the rule-of-thumb described earlier (only voting if the existing value on the post is comparable to the vote that they have). But I suspect the reality is that there aren't many of these kinds of people to begin with, and most people will vote on what they like. (After all, they still give 75% of their vote value to the author).

Which brings me to one more point: Yes, this applies to comments too. If you're thinking about how curation is funky on a comment, forget about it. The comment poster is getting 75% of what gets put on the comment, so just stop thinking about the curation aspect and vote on comments you like. They aren't worth as much as posts anyway.

Again...

  • You don't need to wait 15 minutes to validate someone with a vote.

Other Observations

Another consequence of this setup is that there's one good strategy that I think is a good incentive for us all, and that is the following:

  • If you see an undiscovered post that is good, place your vote on it (at or greater than the 15 minute mark), and then go court a curation group to bless it with a larger vote. You will reap good curation rewards this way.

That also incentivizes people to take part of curation groups because they themselves can front-run a large curator vote. Again, remember that just front running the large curator vote doesn't guarantee anything. It depends on what value is assigned to the post before your vote too!

Summary

So let's summarize what we've learned:

  • Vote before the hump, but not too early.
  • You don't need to wait 15 minutes to validate someone with a vote.
  • Find an undiscovered gem, then court a whale (curation group).

Bonus

If you got this far, great! I have something interesting you might want to play with. This is a curation analysis tool that I used to adjust auto-voters (because, people set auto-voters on the folks they know they like anyway). But it also helps community voters decide when to place votes too, and so I used it to analyze when @helpie should place its vote. Here is what the tool looks like:

Curation Analysis Tool

As you can see, you put in the voter, and optionally the author (you can leave this blank), and it generates a per-post report on where you voted, and where you should have voted (assuming other voters didn't change) assuming the same vote value. And at the bottom of the page, there is a summary:

Curation Analysis Tool Summary

This gives a recommendation for auto-voting time, which is a weighted average of the optimal times in the chart above. You may need to spot check the table to figure out if there were special circumstances for some posts and make a decision separately.

Feel free to play around with it and leave suggestions, and feel free to ask any questions below. Cheers!

And if you like it, please consider a vote for @helpie for our community witness!

vote for helpie

Sort:  

Nicely explained!

And that also means that if you're looking at a post whose current value is way more than what your vote value is worth (...) You ain't getting any curation anyway.

Not true.
If someone votes after the 15th minute, even if the post already has $1000, he will get curation rewards equal (at least) to 1/8 of his vote.
https://steemit.com/steemit/@jga/how-to-optimize-your-upvote-mathematical-analysis

Ah, thanks for the reference! Funnily enough, I had a similar derivation but didn't realize it was saying exactly that (was looking at absolute fraction of curation instead of vs own vote). Will correct!

Posted using Partiko Android

Well, I'm such a little doggy that I mostly just vote what I like. I have pushed my voting trail out to 20 minutes but that is my only concession to the new way.

I used to vote 0 minutes, because I knew the author got it all. I actually vote support for people rather than curation rewards.

Thanks for a great explanation and when I have enough SP to make any sort of difference I'll get in line.

Is it 20 because you want to give the author time to self vote before hand? Otherwise 15 will be strictly better for you.

I pretty much just vote what I like too. I generally don't like to wait around for 15 minutes or so to roll around.

i am a bit slow so bare with me!

Till now i manually cast my upvotes to the ones i am following, sometimes i may check some posts through some resteems and when i have a bit more free time i once again search manually for hidden gems!

i used to do that a lot, spending hours and hours but now due to lack of free time this happens only a few times

I haven't completely understand that part

and then go court a curation group to bless it with a larger vote. You will reap good curation rewards this way.

you mean if we find a hidden gem to go to to a group-community whatever in discord and tell them "here you go i found that one check him and upvote him?

Most of the groups don't they have curators and post promote channels for that reason? i don't mind doing that but will it be ok or they will just ignore me or reply "we have already curators" (they may even be biased that i know that person etc etc)

Well, I suppose what I said is easier said than done. However, if you find a truly remarkable post, it shouldn't be hard to convince curators to vote on it. But you really need to sell it because most people are just dumping links everywhere ... so most just get ignored. I myself don't do this often because it's hard to find really stand-out posts. If we all knew how to do that, we would be a curator for Curie ;)

i don't think it's that hard to find an awesome post in terms of "you have to have an eye for this" and regarding the "curator for curie"

It's just a matter of how much are you willing to spend to find because the whole manually search the way it is makes it hard. Who is gonna put himself though in the process of finding a great post (even if he just finds it in no time randomly) and then start persuading and "selling" it to some curator or curator's groups.

It's tiring, takes once again too much time, he/she may put himself/herself in a "debate", he/she will earn nothing, so at the end of the day for a random person to do all of this is more likely not gonna happen. Although if that post is so magnificent or a post of your friend then you may try and do it with ease.

sorry for the late reply

Wow great info, as I have never paid attention to when or what time I'm voting or commenting......mmmmm how does anyone have time to pay attention to all of that, I get it about the bots though, that stinks! I just look at all kinds of stuffs on this platform.. and engage meeting cool folks, learning and being entertained daily, sooo go votes, go curation go STEEMIANS and thank you @helpie..... I better unproxy @yidneth so I can vote @helpie for a witness!!

Ya, no need to worry about it, pretty much :)

Awesome summary, @eonwarped. I had once post, which paid only approx. 3 % to curators.

Ah, interesting information and it is useful to know how the curation rewards work, but I'm too lazy to try and maximise my curation rewards... I just vote when I read it...

am using the 30 min on my voting list it sems to be working. and a had way better luck on small good ones then the big bot post. I did some research on it.
a like the plan on vote before the hump :)

No reason to use 30 any more, may as well move them to 15, they moved the window since HF20

Thanks to @eonwarped for sharing your observations. :) I am excited about curation analysis tool. I didn't know it before.
I have a question. What are your experiences if "Optimized Upvote Time Recommendation" has the decimal fraction like 9.5. The number shall be rounded off to the next higher or lower whole number?

Probably lower, since you want to front run other votes. Let me know how it goes :)

Posted using Partiko Android

My brain is bleeding, but that's because I'm reading this too late at night. Thanks for explaining this so expertly! Definitely one I'll need to come back to.

To listen to the audio version of this article click on the play image.

Brought to you by @tts. If you find it useful please consider upvoting this reply.

Coin Marketplace

STEEM 0.36
TRX 0.12
JST 0.040
BTC 70446.49
ETH 3571.68
USDT 1.00
SBD 4.73