Bidbot Experiment: The Good, The Bad, The Ugly

Readers who have been with me awhile know I ran a few Steemit bidbot experiments last year, but they were small experiments. Small bids, small amounts, and small rewards. A couple of times, I'm sure I lost money on the bots, though I didn't count too closely. The idea was to merely get familiar with them so I could understand what the issue was with the bots and get some perspective on why they've been disruptive to the platform. My curiosity satisfied, I quit playing with them after a week or so.

A recent controversy surrounding allegations of vote buying in another context piqued my interest again in the bidbots. What could be so wrong with the process of increasing one's rewards with a small investment intended for just that purpose? After all, isn't that what Wall Street investors do every day?

I think the situation is a bit more nuanced than that, and I hope this post illustrates just how.

Image from Pixabay.

My Latest Steembot Experiment

So I decided to run a new Steembot experiment. On Sunday, March 17, 2019, at approximately 9:44 p.m., I published a post titled Experimental Post. It has absolutely no value. It's a completely useless shitpost, published for one purpose only: To buy a bidbot vote.

Unlike the majority of my posts, I did not promote it in any way. I published it, tagged it with the #Steemit tag, and that's all I did until I placed my bid three days later. My hypothesis was this: That one post would earn me a greater return on investment (ROI) than an average post would earn me with my normal routine.

So I should pause right here and discuss my usual post publishing and promotion routine. First, I write the best post possible. I put 100% of my energy into writing the best damn post on whatever topic with every single post I write. No particular reason other than personal pride. Secondly, I tag each post with the five most appropriate tags I can think of based on the topic ensuring that one of those tags is #powerhousecreatives. After that, I share the link in more than 20 Discord servers, pausing to read, vote, resteem, and comment on the posts of others as per each server's rules and my personal interests. That process usually takes an hour to an hour-and-a-half and includes automated upvotes in those servers that offer it and for which I qualify. In other words, it's considerable work. When you factor in that I spend no less than an hour, and sometimes up to three (I've been known to spend six), crafting a post then spend an hour, and sometimes up to two, promoting it, I put in some serious time. I did none of that for my experimental post.

Nevertheless, within 14 minutes, I was able to garner 19 votes for a total of $.23 STU, an humble return. In nine hours, that number was 71 votes for a total of $1.03 STU. In 24 hours, I'd received 76 votes for a total of $1.05 STU. And 24 hours later, no extra votes.

So, within 48 hours, I'd received a potential payout of $1.05 STU with no promotion whatsoever, and I had not yet made my bidbot investment.

On March 20, 2019, at approximately 1:15 p.m., I placed a 40 STEEM bid on The Rising bidbot. You can see the value of that bid, in the image below, was $19.48 USD.

Interestingly, right about that time, I received three upvotes from @up1, @lereve, and @antonella. I have no idea if there is a connection, but it's an interesting correlation.

These three upvotes took my total upvotes, before the bidbot vote, to 82 for a total amount of $1.19 STU.

Within the hour, I'd say about 30 to 40 minutes later, I saw the return on my investment (ROI). After The Rising bot placed its vote, I had 83 upvotes for a total of $27.81 STU.

At first glance, it looks as if I've lost money, but these numbers don't tell the whole story. Let's calculate the ROI.

Steembot Lottery Winner

The way Steembots work is quite interesting. If you go to Steembot Tracker, you'll find a list of them.

It's a rolling list, which means the bot at the top of the list is the one that votes next. You have to get your bid in before the next vote. But the most important information is on the left side of that grid.

The vote value is how much a 100% upvote from that bot is valued at the time of the upvote. On the list above, the top bot on the list has a vote value of $42.81, but that's not the vote value that everyone who bids on the bot will get. Rather, the value of each upvote is based on how many people bid on that bot and the percentage of the total bid that each contains. Let me illustrate by example.

Let's say XYZ Bidbot offers an upvote value of $1.00. If one person bids $.25 and no one else bids on that bot, then that one person will get the full $1.00 upvote. That's a good investment. But let's say two people bid on that bot; one of them bids $.10 and the other one bids $.40. In that case, the total bid was $.50, however, the first bidder registered 20% of that bid while the other tallied 80%. Therefore, the first bidder receives a $.20 upvote while the second bidder receives a $.80 upvote, totallying $1.00.

To see how many other bids there are on a bot before you place your bid, you have to click on the orange Action button and then click on "Details." It's tricky figuring out how much is too much of a bid.

At any rate, how much ROI did my 40 STEEM bid net me?

At the time of my bid, STEEM was listed at $.490176 USD at CoinMarketCap. That means I effectively threw $19.60704 USD into my investment. At the time of the actual vote, the price of STEEM hadn't changed much, so I'll use these figures as I continue (This is a good reason, BTW, to cast your bid as close to time of the next vote as you can).

It's important to point out that the default reward ratio on payout is 50% Steem Power and 50% SBD, but the author only gets 75% of payout while all the upvoters (including the bot) splits the rest. Yesterday, this post paid out and I received $20.71 in author rewards.

Here it is in my wallet:

At the time of payout, SBD was at $1.05 USD. That's equivalent to $10.87485 USD. Since STEEM was valued at .475972 at that time, my Steem Power (SP) payout amounted to the equivalent of $10.629406704 USD. Add those two figures together and that comes to $21.504256704 USD.

Again, my initial investment was $19.60704 USD. Subtract that from my earnings, and my shitpost ROI was $1.897216704. Subtract the $1.19 STU in upvotes I got prior to my investment (valued at .58330944 USD) and my ROI from the bidbot turned out to be a measly $1.313907264. But imagine doing that five to ten times a day every day for a year. The potential return is enormous.

Contrast these earnings, however, with the next post I wrote, going through my normal routine, two hours before payout and I sit with $2.38 STU for payout. STEEM currently sits at .453335 USD and Steem Dollars at $1.02 at Coinmarketcap. Taking 75% of the STU and dividing it in half, multiplying both SP and STEEM projected payouts means my ROI for that post, with about two hours of time invested, comes to $1.314965875. Trust me, my time is worth a hell of a lot more than that.

What's Quality Got to Do With It?

It makes me wonder, if I can earn as much ROI from a single bidbot vote, which takes all of five minutes, as I can from a one hour investment in promotion, then why shouldn't I? So why didn't I experiment with my most recent post instead of creating a shitpost? The answer is, I needed a control, a stark contrast to my usual posting and promotion routine, in order to isolate the results.

From this experiment, I can ascertain that, had I invested in the bidbot vote for a normal, average post, my ROI would have been very similar. Posting the Steem Monsters post the very next day and waiting until payout to post again means other factors affecting payouts were limited, allowing me to get a realistic look at the bidbot effect. So why does that matter?

It matters because this very average post of mine has earned a total of $2.38 STU with only 47 upvotes in seven days.

I can assume that the people who voted for that post thought it worth their percentages. Otherwise, they would not have cast their votes. The exception to this would be those kind folks who auto-upvote every post of mine simply because I impressed them at some point in the past. Those people also voted for the shitpost, by the way. But if you examine all the upvotes on both posts, you'll find there are some voting accounts on each post that you won't find on the other. These are the manual human upvotes. These are the upvotes on which quality--actual post quality--ought to be measured. If we were to put these two posts side by side, we could judge the average power of manual upvotes for each of them. On that, we could ascertain the relative value that each post contributes to the blockchain.

As you can see, the experimental post has a higher total upvote value, due in large part to a single upvote by @Neoxian, an Orca with a stellar reputation on the blockchain, who gave it a very generous .56 upvote. For whatever reason. However, as I hypothesized, the Steem Monsters post garnered a higher average among the manual human upvotes.

How did I determine what is a manual human upvote? The process was simple, but it took a little time. There are some accounts that I know follow me and upvote every post, no matter what. Then there are accounts that upvote me based on a membership. Included in that category are curation trails and Discord channels such as Power House Creatives (aka @steemitbloggers), @qurator, and @steembasicincome. A few of the Discord channels have an upvote bot that will upvote specific types of posts or offer free upvotes based on other criteria; examples include @steemmonsters, which will upvote Steem Monsters-related posts, @esteemapp, which only upvotes posts made from the eSteem app, and PAL (Peace, Abundance, Liberty), whose criteria includes a 30-hour cooldown period. Finally, I compared the two posts and some other recent posts and removed those accounts that were common, assuming them to be auto-upvotes. I know I have quite a few of these because I usually can get more than $1.00 STU in upvotes before I can finish my normal routine of sharing in the various Discord channels I frequent.

All of this is to say that quality is a huge determining factor in what kind of rewards are to be expected. However, quality alone won't do it. One has to put forth the time and effort to promote one's posts no matter what kind of quality sits on the table. Without promotion, results will be meager at best.

NOTE: Using my normal posting and publishing routine, I have managed to attract many @curie votes in the past year, pushing those posts that receive such votes considerably higher in ROI. This further illustrates the importance of quality since Curie doesn't upvote posts promoted by bidbots.

What Bidbots Do to the Steemit Rewards Pool

There's just one more thing to evaluate with regard to Steemit bidbot voting. What does it do to the rewards pool?

Like bidbot payouts, the blockchain creates a fixed number of potential rewards each day. Those rewards are divvied up to all the parties with a vested interest in the blockchain. Without getting into the actual numbers involved, I'll illustrate the principle with an analogy.

Let's assume the blockchain creates 40,000 STEEM on any given day. 75% of that, or 30,000, goes to authors and curators. Let's assume that two-thirds (20,000 STEEM) of that goes to authors and the remaining one-third (10,000 STEEM) goes to curators, the category into which bidbots fall.

If we divvy up 10,000 STEEM among all the curators from redfish to whale, including bidbots, each curator will receive a proportion of those rewards based on the strength of their individual SP. An example of a bidbot with relative SP strength is @buildawhale, which currently has over 67,000 SP and another 1 million+ delegated to it. With a rank of Orca III, it's closing in on whale status pretty quickly.

@therising, the bidbot I used for this experiment, is in the same status - Orca III. It has over 59,000 STEEM in its wallet and an additional 3 million+ delegated to it. It also claims to share 100% of its bot earnings with delegators, which likely explains why it has so much in delegations. In my view, that's a plus.

All of that aside, however, taking another look at the rewards from my experimental post, curators earned $6.74 STU from that post. The lion's share of that went to the bot, which contributed $26.27 STU of the total upvotes for that post. Had I not bid on the bot, that would mean the other curators would have had to split $1.18 STU.

Let's do some more math: The bot's portion of the total upvote value for that post equates to 95.7%. If the bot received that percentage of the total curation rewards, then it took $6.45018 and left about .29 STU for everyone else, a paltry sum to divide between 25 accounts. On the other hand, those 25 accounts would have divided $1.18 STU without the bidbot vote, the lion's share of that going to @neoxian, a manual human voter.

So what do bidbots do to the rewards pool? They sink their teeth into it and rip it to shreds. And this is why you hear so many people, like @quillfire, constantly haranguing witnesses and whales for supporting the bots. It could also be a contributing factor to why there was such a precipitous decline in user activity last year (The decline in STEEM value during last year's bear market was also likely a contributing factor).

From

@arcange: Steem Statistics – 2018.12.25

The Residual Effects of Bitbots on Steemit Post Quality

So far, I've shown that bidbot upvoting can lead to higher author rewards and lower curator rewards for individual posts, the effect of post quality and promotion on manual human upvotes, the difference in ROI between the two, and the negative effect that bidbots have on the rewards pool. But do they have any effect on post quality? I would answer in the affirmative.

Let's take a sample post:

This two-day-old promo post on EpicDice has garnered $263.09 STU from 682 upvotes. That's about the number of votes I've been known to receive from a @Curie upvote that earned me in the range of $40 to $60 STU. Granted, the post isn't nearly the shitpost that my experimental post was, but it isn't any better quality (and probably slightly less) than my Steem Monsters post either.

@EpicDice only has 69 followers and is barely one month old. I would wager that a good number of its manual human upvotes come from its Trending status as a result of its multiple bidbot investments. And the Trending page is full of such examples.

EpicDice is an account worth over $1,000 now. Not bad for only four published posts. Clearly, this account started with someone's personal investment, but it also sits with over 20,000 SP delegations. I won't claim it doesn't add value to the blockchain. There are people who love to gamble, though I'm not one of them, so providing a source of entertainment where gamblers can enjoy themselves over a game of chance does offer some value to some people. Nevertheless, new users should not visit the trending page only to find half of one account's published posts trending because that account bought its way there.

The Steem blockchain is based on a single principle: To reward participants based on the value they bring to the blockchain. We must realize this value is relative. It isn't objective. That's why rewards are determined in part on an upvote system. All participants have the opportunity to vote their subjective judgment on the value of others' contributions. The practice of bidbot upvoting, however, subverts that.

I'll state for the record that I've got no issue with the bidbots themselves, or for their existence. In a free market economy, and the Steem blockchain is a free market economy, one should expect some choices to devalue the community standard. Wall Street has its junk bonds, but it also has rules to regulate bad behavior.

I applaud those efforts to promote the blockchain and Steemit in order to attract new users. I also applaud the efforts of communities such as Power House Creatives and Qurator to promote quality above crass capitalism. But if the efforts of fine folks like @raj808 to get Steem listed on Coinbase and other exchanges is going to have any real long-term effect, it's going to require cleaning out the pool (and I don't mean the rewards pool). We can start with the trending page. What kind of effort would it take to create an algorithm that diminishes the influence of posts that rely heavily on bidbot upvotes? Not much.

Disclaimer: Any error in math is surely my fault. I do my best, but I'm a wordmith, not a mathematician. Also, my apologies to any curators negatively impacted by this experiment. I hope you can find it in your heart to forgive me.

Get your weird lit on:

The Biblical Legends Anthology Series