Steemit Payout Improvement Suggestion

in #steemit5 years ago

Instead of having Steemit community members vote by clicking on a button, something that requires virtually no effort/ time on the member's part, have them vote with their time*.

Yes, Make the "Time Spent on Post" One, of Two, Weighing Factors on Voting Power (the Other, Of Course, Being Steem Power)

Why?

With Steemit's current voting system it's all too easy and, IMO, far too common for people to up-vote a post without even taking the time to click on the post's link to have a look at its content. People are simply gaming a post's "perceived popularity", without so much as giving a minute of their own time to really rate it. It literally takes less than two seconds to up-vote a post and there's little incentive to up-vote content that we really like, because everything (in terms of financial rewards) is based on what up-buttons we click on.

Now, if we were to suddenly make it where each member votes with their time spent on a post and the accumulated time spent on a post between all members were the determining factor of author payout, these members would have the added weight of now having to spend their valuable time and energy to cast their "likes". IMO, people now would be less likely to vote on content that they really don't like, because doing so would require spending more time on posts that don't really interest them.

Of course, having more SP should still offer more curation rewards and it can, with a simple formula which takes both time spent on a post and SP into account.

Example of How this System Could be Done:

I'm going to use a few arbitrary values as an example of how this could work in the Steemit "real world":

Let's say that SP holders are rewarded an additional one one thousandth (1/1000) of a second for every 10 seconds spent on a post, per every 1 SP. Now let's cap the time spent at a 30 minute limit, because. let's face it, a blog post should never take more than a half of an hour to read through, even when it's relatively long.

Why the caps on voting seconds per post? Please take a peak at the image below and then I'll go into it a little later on.


The above image is taken from Steemd.com and it lays out the current SP distribution of Steemit's members (where 1 MV = 1M VESTS = 339.867 STEEM = $180.469). As you can see, 51.27% of Steemit accounts are "newbies", or have SP values worth around $1.80 (somewhere between $0.01 and $10, I assume), yet they only combine for 0.02% total stake in SP.

This means that if every account up-voted one time in a 24 hour period, each on a different post, each being the only up-vote for that post, and assuming that they all average the same number of votes per 24 hours (voting power calculation takes both SP and votes per hour into account), then the 1,944 newbie accounts would only distribute 0.02% of future Steem. This explains why a minnow's early up-votes (first up-votes to a post) are only worth fractions of a fraction of a penny to the post author at this time.

Now, Consider Adding in the Element of Time to Weighing the Voting Power of Any Steemit Account

Let's say that we use the following simple formula, as an example:

(T/ I) * (S * V)

Where;

T = time an account spends on a post, in seconds (capped at. let's say, 30 minutes worth of seconds = 1,800 seconds, payable)
I = "voting second amplify time", number of 10 second intervals that the account spends viewing the post
S = "voting seconds" (VS) payout of, let's say, 1/1000 of additional seconds for every 10 seconds spent on post, per 1 SP.
V = account SP value

Using the above values, a 30,000 SP account spending 15 minutes on a post would provide the following payout:

[(15min * (60s/ min)) / (10 seconds voting second amplify time)] * [(1/1000 voting seconds per 1 SP) * (30,000SP)] = 90 * 30 = 2,700 additional voting seconds

We'd add this value to the "actual time spent on the post" (15 minutes, or 900 seconds) to get our total payable, curation reward, time of: 900s + 2,700VS = 3,600 "payable curation time" (the value that's used to calculate the percentage of Steem payout to the content author and the potential curation reward payout to the account which made the vote).

The "actual time spent on the post" (ATOP) can be divided by some factor to decrease the weight of that value on curation reward potential. For instance, we can divide ATOP by 10 to give that value 10 times less weight compared to SP.

The above example would then have a total payable curation reward time of:

(900s / 10) + 2,700VS = 90s + 2,700VS = 2,790 payable curation time.

In fact, we can give time spent on the post any weight compared to SP that we desire, either choosing to multiply or divide that value by any factor and keeping the SP calculation (voting seconds) the same.

This is where we can give minnows more voting power. Let's assume that we stick with the original calculation (without diminishing the weight of ATOP), keeping the SP portion of the calculation as it is:

Now, let's look at a minnow that has only 10 SP, but spends 30 minutes reading a post. Here's what the voting power calculation would look like:

[(30min * (60s/ min)) / (10 seconds voting second amplify time)] * [(1/1000 voting seconds per 1 SP) * (10SP)] = 180 * 0.1 = 1.8 additional voting seconds

then, 1,800s + 1.8VS = 1,800s + 1.8VS = 1,801.8 payable curation time.

Now let's compare the two examples.

The 10 SP account spent twice the time on the post as the 30,000 SP account. This is time that could have been spent reading another post or doing something else - time is valuable and so is the energy that's spent by our computers as we keep a post up inside of our browsers (therefore we're punished for keeping a post open that we don't really like)! So the lower SP account spends twice the time in order to provide the content author with about half of the Steem reward and, therefore, has half of the potential curation reward payout as the higher SP account, all else being equal.

How About the Whales?

Now let's consider an account that holds over 1 million SP. Let's assume that this account spends one minute on a post. Here's what the calculation would look like:

[(1min * (60s/ min)) / (10 seconds voting second amplify time)] * [(1/1000 voting seconds per 1 SP) * (1,000,000SP)] = 10 * 1,000 = 10,000 additional voting seconds

then, 60s + 10,000VS = 10,060 payable curation time.

Now let's compare this whale vote to the 30,000 SP account vote and the 10 SP account vote.

This whale spent 1 fifteenth the time of the 30,000 SP account and 1 thirtieth the time of the 10 SP account on "voting", yet still provides a 279% and 558% better author reward and curation reward potential, respectively. The whale's much higher SP value makes his/her time more valuable to the post author and rewarding to the whale, in terms of curation rewards.

The main difference between this approach to distributing Steem versus the current model is that it should work out where the whale has significantly less voting power compared to minnows, unless, and here's the rub, the whale decides to spend a lot of time on a post(s).

Of course, it would have to be worked out how to make it such that one account can only be spending "voting time" (*VT) on one post at a time and, IMO, there should be a cap on how much VT any account can give to a post. Otherwise we could have a situation wherein a whale leaves a post open for days on end and the post author gets an endless stream of high Steem payouts (or, at least until the payout cut-off).

Let's Say that We Cap VT at 30 Minutes for Every Account

Now a whale can pay an author up to 30 minutes worth of TSOP.

Going back to the 1 million SP example, the capped limit would look as follows:

[(30min * (60s/ min)) / (10 seconds voting second amplify time)] * [(1/1000 voting seconds per 1 SP) * (1,000,000SP)] = 180 * 1,000 = 180,000 additional voting seconds

then, 1,800s + 180,000VS = 181,800 payable curation time.

Even if the whale keeps the post up for 300 days, his payout to the author will remain capped at 181,800 "payable seconds" of vote, and his curation potential will remain capped there as well.

I'm not convinced that any accounts should have a cap on how much voting power that their time spent on any new posts will have, even if they've reached the time limit on prior posts. This way, our time spent on Steemit is highly incentivized.

I think that the current curation system of paying out the majority of the supply to the "early voters" should remain in order to battle against people jumping on late to bank on all the voting hours that go into a post. I also think that it would be beneficial to delay the reporting of voting hours by an hour or more in order to hide whale votes long enough to make the high payout content (in terms of curation rewards) less noticeable.

Everything should be geared at diverting most of the user bases' focus at spending more time on content that they actually like, versus on those posts that they believe will be more popular with the majority of the user crowd.

Now Let's Zoom and Take a Look at the Whole Picture Now

So each account has an attached SP value. Most accounts (over 50%) have between 1 cent and $10 worth of SP. That being said, all of our time (all Steemit users) is valuable.

My suggested payout system takes the second point into account. Instead of basing Steem distribution on where we each place a few clicks, my system bases it on where we each spend our time and it only rewards those that take the time to look over Steemit content.

Therefore, if the Steemit whales spend a combined 30 minutes worth of time looking over Steemit content within any given 24 hour window, then they will only be rewarded for those 30 minutes worth of VT and they will only take up that portion of the total VT pie for the day. Should the minnows combine to spend hundreds of thousands of hours looking over posts, then their combined time and energy will more than make up for the whale's higher SP, and their votes will make up the majority of the voting power for that day as a result.

This is easily proven by looking at the math:

Here's for 30 minutes of whale VT, averaging 1.5 million SP:

[(30min * (60s/ min)) / (10 seconds voting second amplify time)] * [(1/1000 voting seconds per 1 SP) * (1,500,000SP)] = 180 * 1,500 = 270,000 additional voting seconds

then, 1,800s + 270,000VS = ** 271,800 payable curation time**.

Here's for 1,000 accounts, of 10 SP each, spending an average of 2 hours reading over content:

[(120min * (60s/ min)) / (10 seconds voting second amplify time)] * [(1/1000 voting seconds per 1 SP) * (10SP)] = 720 * 0.01 = 7.2 additional voting seconds

then, 7,200s + 7.2VS = ** 7,207.2 payable curation time** per minnow.

therefore, 7,207.2 payout per minnow * 1,000 minnows = 7,207,200 total payable curation time for the 1,000 small SP accounts

In this case, assuming all else is equal, the minnows would provide a combined 2,669% more voting power than the whales. Why? Because, yes, SP is valuable, but time is valuable, too. Thirty minutes of combined whale up-voting, no matter how high their combined SP values, should never be worth more than 2,000 combined hours of up-voting, I don't care how low their combined SP values.

This is what Steemit is missing, IMO. Our time spent curating needs to be rewarded, not just our SP value and the timing of our up-votes.

Note Also...

That the cap limit doesn't really limit how much a post can potentially be paid out on any given day. If, for instance, every active account for the day were to spend all their time viewing a single post, be it for only a few seconds or for well beyond 30 minutes on average, then all the Steem payouts for that day will only go to that post author. It all comes down to the pie chart of VT distribution for the day, taking the 30 minute payout limit into account.

Furthermore, poor quality posts are very unlikely to receive high Steem payouts, since the combined VT that goes into their post would have to take up a relatively large portion of the combined VT pie chart for the day in order to receive a high percentage of the daily SP payout. Even having 30 minnows vote with their time, each spending 1,800 seconds on a post, would only amount to about 54,000 payable time, which is easily surmounted by a single whale spending a minute or two on some other posts. The difference now is that whales are forced to spend more than a few seconds to vote up posts; they're forced to actually look at the content!

It would seem to me that this would help out with the bot system that seems to be taking over on the curation side of Steemit over the last weeks, but perhaps I'm missing something (could the bots be designed to open posts from popular authors and then spend 30 minutes on them before moving on to the next high payout potential post?) Any feedback on this would be appreciated.

I Really Think this Suggestion Could Improve Steem Distribution and Bring More "Fairness" Into Steemit's Up-Vote System, Particularly from the Minnow's Perspective...

But I'm aware that I can't see all the angles on this and I don't know how this kind of system could be programed into Steemit, or if it's even possible.

That's why I would like to create a discussion about this suggestion with anyone who's willing, @jamesbrown. I invite constructive criticism, tweaks to my suggestions and alternate solutions. Let me know if what I suggest is feasible and/or if it's even an improvement. Share why it can or can't be implemented into code, if you're skilled in that area and happen to know if, and how, it can be done.

Thanks for your time :)

Sort:  

This is interesting, but it is really just the voting target with a couple of extra steps.

SO for example, lets say we set the amount of time required for a full power vote at 30 minutes, this would mean we could cast 48 votes per day at 100%, 96 at 50% etc etc.. this is the same operational effect as the vote target effecting voting power.

It would have the effect of preventing an inactvie user from casting his unused votes randomly (though i don't necessarily see that as a problem right now.)

But it would also penalize the pickier users.

I kinda like this idea, something that could be fair to it would be this.

You open post, your voting button shows a percentage starting at 5 or 10 slowly going upwards depending on time and the time it takes the average reader to read the amount of content and words in the blog.

Thus they can still vote, but if they decide to stop reading midway their vote wouldn't give them full power thus decreasing the curation rewards to it.

Things could still be tweaked etc, but I am not sure its something Steemit would want. With the decentralization and freedom of users.

I like it! Perhaps they could also be given the option to "take back the vote" after having time to read and digest the content. This way they're given the option to only vote-up the content that they really enjoy and appreciate.

Seems fair enough to me. Whales can still have the majority of voting power if they want, and are willing to put in the extra effort to look over some posts. Time being rewarded makes sense. It also makes sense that the post receiving the most accumulated viewing time should get the most rewards.

I'm not sure about the tech side and how one would go about implementing the ideas into code, though, nor the specifics of what the values in your formula should be (maybe "I" in your formula would work better at 30 seconds or 1 minute? This would likely have to be tested in beta before going live.)

Yours are my thoughts, exactly.

As far as your comment about the formula values: I agree that they can be calibrated to be more efficient. The I value in particular might need to be tweaked. Perhaps it could be determined based on like the 3rd or 4th standard deviation of the mean average user time spent on a post.

Let's say that the average works out to about 2 minutes and 30 seconds (or 150 seconds) with a standard deviation of about 40 seconds, then post viewing outside of 3 standard deviations (below the mean) would come in at about 150s - (40s * 3) = 150s - 120s = 30s. So we could attempt by starting at an I of 30 seconds. This would make for up to 60 paid intervals within the 30 minute limit, per post voted on, compared to the 180 paid intervals at an I of 10 seconds.

It seems very ironic that I up-voted this post...

LOL.

Ironic or not, this is a topic (Steem distribution/ voting model) that needs serious discussion, IMO.

People will intentionally start to write longer posts. The longer the post the less I want to read it. This means that I will probably skip the post without upvoting. Congrats! Steemit is now full of posts that no one wants to read.

If an author puts out bad content, long or short, people will be less likely to read his/ her future content. It's like with anything else, if you like it then you're willing to spend more time with it (on it). Additionally, a feature that allows you to "take back your voting time" would remove any rewards given to the author based on their lengthy article.

Also, it might also encourage posters to cover their topics more thoroughly, making for more valuable content.

Actually they have something in place now, it's kinda like a game, somewhere in the first 30 minutes of a post is the magic minute to cast a vote. So it's different everytime. or so I have read. Still confuses me how it all works. This steemit is like a game. Got to figure out how to play it. Good post :)

That's another point for another post - this confusion about how voting really works.

But the problem is people trying to game the system by jumping ship to the next post that a whale votes on and people basically following these limited accounts with high SP. This leads to a lack diversity of SP distribution to the few accounts that the whales game for curation rewards, such as @dollarvigilante.

If people were to instead get rewarded for the amount of time spent on their post(s), then I'm certain that @dollarvigilante would get far fewer rewards per payout, because probably more than half the people that vote on his content don't even care to read what he has to share (they're just gaming it for curation rewards). That's not to say that @dollarvigilante 's content isn't high quality, but, rather, that too many accounts are voting purely on popularity.

If we could somehow build a system that better incentivizes reading content that we actually enjoy reading versus doing so for "popular" content and, at the same time, makes it more difficult for whales to set up "lazy bot systems" that mindlessly vote for that popular content and then get the hive following of votes behind them, then this should make for an all-around fairer and more enjoyable experience. I think that this "voting with time" idea is a step in that direction and it just makes sense - time is valuable and where/ how we spend most of our time says everything about what we truly value.

Can I upvote at 4 days?