Introduction
A few months ago, I began a series of articles that was intended to make the case that not only is automated voting unavoidable in the steemit ecosystem, it is also desirable. In the first article, Maslow's Hierarchy of Needs and The SteemBot Revolution, I argued that automated (bot) voting is guided by human designers, and is thus constrained by Maslow's hierarchy of needs - just like the humans who participate in steemit's ecosystem. In the second, On SteemBots and Voting Errors, I borrowed from statistics and introduced the concept of Type-I and Type-II voting errors. In this framework, a Type-I voting error occurs when I vote for an article that is not actually valuable to me and a Type-II error occurs when an article that I value does not get a vote that I could have cast.
[Image Source: pixabay.com, Licence: CC0, Public Domain]
I had an insight on voting the other day in the comments of a post by @dwinblood, where @dantheman said:
I view down vote as up voting everyone else, but the downvoted item, just more efficient. Every upvote implicitly reduces rewards of everyone else.
It's the old Abbot and Costello routine. The Drill Sergeant asks for a volunteer to step forward, and everyone but Lou Costello steps back, so Lou is left standing out as the volunteer. The downvote (and the upvote) can be viewed the same way.
[Image By Anonymous - http://www.ebay.com/itm/BUD-ABBOTT-LOU-COSTELLO-PUBLICITY-PHOTO-for-the-1953-Abbott-and-Costello-Show-/122255313721?hash=item1c76fc0b39:g:FWQAAOSw6DtYSJfe, Public Domain, https://commons.wikimedia.org/w/index.php?curid=53901173]
So, following that logic in today's post, I am going to extend the argument on bot voting in a new direction. The reality is that when I am not voting up to my full potential, I have delegated my voting influence to everyone else - bots and all. Not only am I committing Type-II voting errors by allowing valuable posts to go by without my vote, but I am also committing Type-I voting errors by delegating my voting influence to people who cast votes that I disagree with.
It's All About Responsible Curating
In a more recent post, Did you give your steemit voting keys away?, @beanz wrote:
Curating takes up a lot of time and attention. When the option to make curation rewards without putting in much effort is there, we can become irresponsible curators. Curating is a big responsibility. Rewarding as many users as possible without making people after months of creating content feel left out can be tough, and it's an importort (sic) part of user retention.
I agree with almost every word of that excerpt. There may be nothing more important to steemit's success than the quality of curation on the platform. There's probably a reason why curation was listed first in this excerpt from the steem whitepaper
However, I think that @beanz is overlooking the impact that not voting at all has on the morale of authors in steemit's ecosystem. When I choose not to vote, everyone else is not standing still. If no one voted at all, the harm would be limited to Type-II voting errors, where no one receives rewards for valuable content. But like the rest of the line in the Abbot and Costello skit, everyone else is not standing still. If I decline to vote, the rewards pool is still going to be distributed, and my influence is delegated to the accounts that do vote. By delegating my share of the voting influence, part of my "share" of the rewards pool is necessarily going to go to support things that I don't like. I'm making the very Type-I voting errors that I sought to avoid by eschewing bots in the first place.
So I am led to ask, which is better?
- To use the best bot I can get my hands on and delegate my voting to the closest approximation to my own tastes that I can find or create (and continually seek improvement so that the bot will do a progressively improving job of representing my own preferences)? -OR-
- To leave it up to chance and delegate my voting influence to all the other bots and humans who will guide the reward distribution?
Conclusion
The voting system has not been clearly described in any recent publication that I've seen, but it is said to be optimized for about 40 votes per account per day. Are you casting that many manual votes? If not, then you are standing still while all the other voters move ahead. Effectively, you have traded your share of curation rewards to the other voters in exchange for having them do the work to distribute the author rewards. Go take a look at steemdb.com and search for your account to see what your voting power is. If it is near to 100%, then understand that you are voting with bots. It's just that you're voting with bots of someone else's choosing, instead of your own.
Two final points: The first is one that I've made so often that it's coming to feel like a cliché, but it is certainly true. Google's PageRank system is nothing more than a sophisticated bot. They take web pages and try to rank them according to humans' subjective tastes. No one would seriously suggest that Google should disable their supercomputers and replace them with human readers. Why, then, do we assume that complex interactions of many steembots can't learn to be just as good as Google at finding quality content for human enjoyment?
And lastly, I am not - by any means - proposing that all curation should be done by bots. Google uses human input in the form of links between pages to guide its algorithms. Similarly, in order to deliver curation that is consistent with human tastes, steembots will always need human input.