The Bot Experience: From a Bot Runner's Perspective

In my time here on Steemit, I started from being anti-vote-bot, to joining auto-bot-voting communities, to using all manner of voting bots, and in the most recent escapade, I helped set up a bid bot for @cryptoempire. And now we've shut the bot down and I'm back to being anti-vote-bot.

The recent discussions concerning bid bots and trending brought me to re-evaluate my position. I agree with the anti-vote-bot proponents in that the bid bot votes eat up too much from the rewards pool while providing relatively little value.

Running the Bot

Our community had been discussing possibilities for a bot to help it grow, and since we were pretty much accepting of bid bots at the time, we thought, Why not a bid bot? And then @spiritualmax thought: Why not a gamified bid bot? Well that was an intriguing idea. The bot account would host a game of sorts, and income from running the game as well as bid income would be distributed to delegators. Delegators would get in-game bonuses.

And getting a bid bot running is quite easy to do, thanks to @yabapmatt's excellent open source code, so we figured we should try it out.

I set up a VPS on vultr for 10$/month, and bought a domain for 12$/year. It's not beefy enough for a witness steemd node from my understanding (only 2G RAM), but good enough to tinker with condenser and good enough to run a bot.

To fund the server and the domain it was running on (for SSL certificate purposes), I took 2.5% of the income from the bot, independent of my own delegation. My rough computations would show this portion of funds would be enough to cover, but the income was highly variable.

The way our bot filled up, the income from delegation is what you might get if you self voted with that same SP about 5 times a day, since our bid rounds were not that full. And that would have been frowned upon if the amount of SP we were talking about was really high.

In times of heavy traffic, the steemd nodes would go all be unstable, and as a result votes and withdrawal payments would fail. This required manual intervention to handle and was quite annoying to resolve. And I’ve refunded a lot of bids manually in this way.

In my opinion, the structure of the bot should be changed so that if such failures occur it should stop everything until the system recovers and the failed vote or payout can be made. I also wonder if there are cases where the API could get interrupted and the refund actually occurred, which would result in a double payment.

Also, in order to be a more responsible bid bot, I was planning to adopt a blacklist as others have done. But since our bot was puny in comparison I did not prioritize it. I had set the settings to still give refunds to blacklisted members out of fairness though, because what's the point of taking money for a service you did not provide? In any case, I only mention this as an added management cost for bot maintenance, making it even less desirable for me to support running one.

The bot is gone, but I’ve kept my server going, as a playground for me to develop side projects.

The interesting thing is that the bot as is could also be configured to just distribute liquid funds the account receives automatically. So if we focused on running a fun game, we could have a similar payout structure. It just wouldn't be as lucrative as a bid bot.

Rationale

There are many good articles bringing up this discussion. But it boils down to the following:

• With SP delegated to a bid bot, the delegator and the bot bidders collectively siphon a portion of the reward pool that is equivalent to the same SP being used to self vote 10 times a day. And most of it goes to the delegator.

The difference is that it spreads the reward to multiple bidders, and multiple posts. This might seem like it's more ideal, and is why I was initially on board, but it misses the larger picture: the distributed reward is not audited at all, and no community feedback enters this picture. That's the same objection many of us have towards self voting.

"Why does there have to be community feedback for reward distribution?" you might ask. Wait, do I really have to answer this question? Fine... Because we all share the same rewards pool, so we should all have a say when it comes to how it should be distributed.

"But we do! It's called voting, genius." Well yes, if we were actually exercising it. But we are not, except in really blatant cases, and even for some of those we can't do anything. How do we say we do not wish bid bot delegators to get so much of the reward pool? We can't. We can only make bid bots less desirable to use.

But what can we do, for those of us that actually care about this? One way we can demonstrate this is to remove witness votes for bid bot supporters.
The other is to exercise flagging on posts to discourage usage. And there are two flavors:

1. Bidders aiming to get small profits.
2. Bidders aiming to get to trending for exposure.

It's quite painful, as bidders are paying for it. If we flag them further, they end up paying even more. But there is no other way. Maybe flag the lousy ones only...

The point is, those of us against this system need to start voting to be heard. Even if we reserve just 1 of our 10 daily votes to flag, the situation can be improved. At the least we start to have community feedback for what is happening.

Note that flags on bid posts have a double effect. It makes bidding less desirable, so delegators to bid bots get less, and the ones that aren't flagged deserve the value.

In the end, witness votes hopefully can change the landscape in the long run. If indeed there are more of us that agree that this is a problem, we can make our voice heard by voting for witnesses that are against bid bots. And come to think of it, why are we rewarding large bot delegators that are witnesses anyway? Aren't they satisfied enough with their bid bot profits? One should ask if it is really good for the platform or not...

The Advertisement Viewpoint

I wanted to take a special section to address folks who are saying that we should view this as paid advertising.

What options does someone have for paying money to advertise a post, if not a bid bot?

Framed in this way, there's no good answer. Promoted section is a joke. Trending clearly has people's attention. Bid bots enable access to trending.

But at what cost? Let's face it, the current advertising by bid bot setup is lousy. Advertisers willingly pay bid bot delegators a large amount for this access. But we should ask, why do bid bot delegators deserve this money? Wouldn't it make more sense for this money to go into the platform itself? And wouldn't it make more sense if it wasn't cutting into the rewards pool? Same question I guess.

And let's not forget, if trending does not have interesting blogs, people will stop going there. I stopped going there long ago. It's also the front page for steemit that outsiders see, and is one of the signals for how they evaluate the platform.

I personally can only come to the conclusion that it is not the right way, and the platform needs to change. Promoted needs to be improved so people can stop talking about bid bots for promotional use.

For now though, it seems there is no alternative. But at least we can reclaim part of the rewards pool, and make the cost for promotion higher. Make bid bots riskier and more costly. Reduce the portion of the rewards pool allocated to bid bot delegations.

See this excellent post about comparisons to a proper advertising model.

Summary

With consideration to the recent discussion, I decided to stop delegating to our own bot. And after that, we had a discussion and decided it made sense to discontinue the bot.

And after that, I removed witness votes (that I know of) related to bid bots. And I decided not to self vote or buy/sell votes.

---

_{Click the banners to learn more about each community!}

---

	Get Universal Income: MannaBase and SwiftDemand
	Trade STEEM on Binance
	Sola.ai :: Earn.com :: Refereum :: STEEM for Tasks :: Sweatcoin