Let's talk about voting bots

Premise: Rewards + Audience is more valuable than rewards alone.

^{pixabay license}

Introduction

During my years on the Steem blockchain, a consistently controversial topic has been the use of voting bots. Without wading too far into the controversy, I'll say that I'm generally in favor of automation, but it's important that it's done well. I'll also say that there are pros and cons for everything, so what I look for is not perfection - but continuous improvement.

For me, it's interesting to look back on how things have changed, and to wonder about what might come next. Informally, I think of the voting bots that have existed here in terms of 4 generations.

Generation 1 Self-managed bots

- These are/were bots, written mostly in python, that let a user automate their own voting with a program that they could run on their own computer (or cloud host). For the most part, these bots used simple criteria like the author and tags to decide what to vote for.

Generation 2 Voting trails

- For people without the computer resources or technical skills to run their own bots, websites and other services began to emerge where a single person would manage a voting bot and also use it to vote on behalf of other users. Typically, the person managing the list votes near the front of the line, which lets them collect extra curation rewards from the votes of all of the other accounts in the list.

Generation 3 Pay per vote

- The first one of these that I remember seeing was @randowhale. In the case of randowhale, the bot-user would submit a Steem post for voting - along with a payment, and the bot would give back a random vote. Later iterations in this generation align the payment size with the vote size.

Generation 4 Delegation bots

- Of course, we're probably all familiar with the current generation of bots that uses investor delegations (and sometimes also pay per vote) and currently dominates the trending list. With these bots, investors delegate to the bot, and they get a daily vote. Commonly, they also get a share of the bot's curation rewards.

_Caveat - I don't have a whole lot of experience with the Gen-3 or Gen-4 voting bots, so this is written from an outsider's perspective, but that's how I think it works.

As far as I can tell, we've been stuck at Gen-4 for three or four years now. So, my problem with the voting bots isn't that they exist. I suspect that they're unavoidable. My concern is that they seem to have stopped innovating. Their investors should demand more.

This morning, I thought I'd try to take an objective look at some pros and cons of the Gen-4 voting bots and imagine what a Gen-5 voting bot might look like.

Gen-4 pros and cons

To me, it seems clear that the argument of whether these bots should exist is pointless. They do exist, and I can't imagine a scenario where they disappear in the near future, so instead of wishing them away, how do we make sure that everyone's interests are aligned? First, we take a clear-eyed look at the advantages and disadvantages, so that's what I'll try to do here (briefly, I am in a hurry today, unfortunately).

Advantages

Immense nominal return for investors: I haven't run the numbers myself, but I've seen claims that bot delegators earn nominal returns on the order of 100% APR. I don't know if that's correct, but they certainly collect sizeable returns.
Because of this immense ROI, the bots probably attract some investors.
The bots certainly incentivize their delegators to HODL powered-up STEEM.
Like it or not, it's proof of a successful business model on the Steem blockchain.

Disadvantages

Earning maximum rewards requires a daily post. I would imagine that most investors would prefer passive returns.
Because daily posts are required, and because the payouts are relatively high when compared to other blockchain content, the trending page is dominated by things that most people probably don't care to read.
This likely drives away audiences that might enjoy reading other content here, if that content were allowed to reach the surface.
Driving away audiences probably devalues the investment stakes that are sustaining the delegation bots. Remember, under advantages, I mentioned large nominal rewards. My suspicion is that the actual returns (after accounting for devaluation of the investor's stake) have been far worse than they might have been if interesting content were allowed to surface.

These lists are far from exhaustive, but since I'm in a hurry, I need to move on. Hopefully, I hit the highlights. So now, let's talk about the potential for a Gen-5 voting bot.

What would a generation 5 voting bot look like?

Off the top of my head, here are some of the characteristics that we'd like to see:

Large passive rewards for investors
Protects and increases the value of the investor's stake
Encourages creation of interesting content
Rewards authors who deliver an audience (i.e. impressions, views)
Does not require, or even encourage, daily posting of vacuous content

Impossible, right? I don't think so. I think that a bot-owner would just need to stitch together some already existing technologies in order to accomplish this. The things that are needed follow:

View counters
Beneficiary rewards
A reward distribution scheme like the ones that are in use by Bitcoin's mining pools.

So, let's imagine a 2-dimensional voting bot that's supporting a stable of authors and also a pool of investors. The authors get paid in proportion to 2 factors: (i) the rewards on their posts; and (ii) the views on their posts. Simplistically, the stable of authors could just be anyone who posts through the bot-owner's web site. (Of course, authors can also be investors, and vice versa, but they don't have to be.)

Similarly, investors also get rewarded in proportion to two factors: (i) the rewards on all posts; and (ii) delegation size.

Now, the bot operator keeps a running tally of how much is owed vs. how much is collected, and they distribute it through the use of their own web site.

The (shall we say) "sponsored" authors post through the bot operator's web site (something like upvu.org). In this way, the bot operator can track their views and set their beneficiary percentages in order to distribute rewards according to the rules above. The reward distribution can be done using something like PPS or PPLNS.

There's some complex math to be done here, but that's what computers are good at, so I don't see it as overly challenging.

If the bot-owner doesn't want to set up their own web site, perhaps they could arrange to get the view count information from Steemit, but they'd need to rely on their authors to set the right beneficiary settings, which is an easy step to forget about.

If they do set-up their own web site, they could also harness the Visibility as a Service concept that I've discussed in the past, which might provide a secondary mechanism to protect and increase the value of the investor's stake.

Conclusion

Gen-1 harnessed automation.
Gen-2 harnessed curation rewards.
Gen-3 harnessed blockchain transfers
Gen-4 harnessed delegations, transfers, and curation rewards

For me, the key insight that Gen-5 voting bots need to harness is this. "author" rewards don't need to go to the author of a particular post. Through the use of beneficiary rewards and reward-sharing algorithms, beneficiary rewards can be redirected to investors or even to other supported authors, and rewards can be distributed in proportion to a healthier mix of factors.

By harnessing beneficiary rewards and audience metrics, I believe that Gen-5 voting bots could profit more and better align themselves with the interests of authors, investors, and audiences.

What do you think?

Update: See also, I Wonder If Bid Bots Can Be Made Less Harmful?, which was posted today by @o1eh.

Thank you for your time and attention.

As a general rule, I up-vote comments that demonstrate "proof of reading".

Steve Palmer is an IT professional with three decades of professional experience in data communications and information systems. He holds a bachelor's degree in mathematics, a master's degree in computer science, and a master's degree in information systems and technology management. He has been awarded 3 US patents.

^{Pixabay license, source}

Reminder

Visit the /promoted page and #burnsteem25 to support the inflation-fighters who are helping to enable decentralized regulation of Steem token supply growth.