Hi Everyone,
I believe this is the post that several of you have been waiting for since I started my game theory series. There are several types of bots. There are the bid-based bots, content promotion bots, resteem bots, subscription-based bots, quality content reward bots, and several other types. The bid-based bots are the most interesting from a game theory perspective.
Everyone on Steemit has some sort of objectives. You could be here to read interesting content, you could be here as an investor, you could here to share content, you could be here to make a quick profit, and you could be here for any number of other different reasons. I’m on Steemit for a combination of reasons. My most important reason is to share content. Using bots to help promote content will be the main focus of this post.
Posts can get buried very quickly in Steemit. As Steemit grows, posts will get buried faster and faster. As it is, within 10 minutes, your post will move so far down the main ‘new’ list, people will not see your post. The use of good tags can definitely help. The popular tags such as ‘life’, ‘photography’, ‘steemit’ and ‘travel’ have a similar problem and the post will quickly be out of view. I mostly use the moderately less popular tag ‘economics’. My post can be near the top of the ‘new’ list for half a day. I imagine that the most relevant tags of most people’s content has the same problem. For the less popular tags, posts can stay on the new lists a lot longer but very few people look at those tags. Posts will still get very little exposure.
Getting on the ‘hot’ lists is important for content exposure. Unfortunately, getting on the ‘hot’ lists is not easy. A post needs to get high value upvotes earlier on. The existence of bots has meant that the value of upvotes needs to be higher to get on these lists. I scrolled down the main ‘hot’ list, I needed to scroll quite far to find any posts older than 5 hours. These posts had a very high upvote value and were also on the ‘trending’ list. The ‘hot’ lists for the tags are similar, I looked at a few popular tags and could not find posts older than 12 hours. The posts that were 7 hours or older were also on the ‘trending’ lists. I looked at my favourite #economics tag, there were several post close to 2 days old.
The ‘new’ lists give minutes of exposure or hours if you use less popular tags. The ‘hot’ lists can give a few hours of exposure or a couple of days if you use less popular tags. So how about the ‘trending’ lists. The main ‘trending’ list can give you a couple of days of exposure, if you have a massive upvote value. The ‘trending’ lists for popular tags can give you exposure for a few days, if your post has a massive upvote value. The ‘trending’ lists for moderately popular tags, such as my favourite #economics tag, can give your post exposure for close to the full 7 days without a massive upvote value.
All of the above is worth considering if you are sharing quality and original content. Most of my posts relate to economics and has limited appeal to the wider audience. My focus is to get my posts on the ‘trending’ lists of the tags most relevant to my content. I have not invested in bots to the extent that my posts get on the main ‘trending’ lists. I would not recommend paying bots to the extent that your posts get on the main ‘trending’ list. This approach might be beneficially in the short-run because of the massive exposure. There are a few users that have been very successful using such a strategy. In the long-run, this strategy will hurt Steemit and ultimately yourself.
Steemit can be seen as a game of prisoner’s dilemma. The more money people pay the bots, the more everyone needs to pay the bots in order to get exposure. This is particularly true of the bid-based bots. The number of bid-based bots have exploded. A few months ago there was a handful. Now there are well over 30. The bid-based bots previously provided a constant but small return on bids. The usage of the bots has mostly drowned that return out as can be seen from the table below.
Returns on investment
| Bid Bots | Sample 1 | Sample 2 | Sample 3 | Sample 4 |
|---|---|---|---|---|
| yourwhale | -23.48% | 50.97% | -7.97% | 219.11% |
| minnowhelper | 99.77% | 84.49% | NA | 520.87 |
| upme | -5.64% | -9.93% | -11.30% | -13.62% |
| kittybot | 69.64% | 28.33% | 2950.65% | 301.06% |
| adriatik | -25.65% | 36.32% | -9.90% | -11.89% |
| appreciator | -6.99% | -9.95% | 1056.66% | 175.53% |
| upgoater | -33.83% | 25.87% | -30.50% | -71.46% |
| minnowvotes | -0.55% | 69.32% | -28.99% | -11.04% |
| levitation | 104.83% | -1.98% | NA | 99.94% |
| aksdwi | 424.47% | 41.28% | 370.63% | 893.39% |
| sneaky-ninja | -10.47% | -10.95% | 19.13% | 14.28% |
| ipromote | -17.06% | 20.16% | 19.61% | -14.81% |
| boomerang | 706.48% | 3.99% | 9.33% | 22.56% |
| upyou | -23.18% | 58.87% | 11.52% | -73.00% |
| sleeplesswhale | -15.99% | 48.21% | -9.02% | 219.31% |
| hellowhale | -28.96% | -16.95% | -10.83% | 4.88% |
| allaz | 2.08% | -8.59% | 6.14% | 305.70% |
| votebuster | -64.14% | 0.34% | -17.09% | 29.76% |
| booster | 171.00% | -2.95% | -27.68% | 129.97% |
| jerrybanfield | 10.99% | 51.40% | 2.27% | -49.91% |
| buildawhale | -14.05% | 10.83% | -9.33% | -14.84% |
| pushup | 43.69% | 2.60% | 11.67% | -5.59% |
| discordia | -70.12% | -69.78% | -66.85% | -92.67% |
| upmyvote | 19.76% | 13.35% | 21.68% | -1.09% |
| steembloggers | -71.29% | 0.66% | -21.55% | 430.70% |
| moneymatchgaming | -34.45% | -68.01% | -46.38% | NA |
| lovejuice | -17.78% | -17.32% | -5.62% | -70.56% |
| upmewhale | -14.45% | 56.16% | -23.02% | -1.29% |
| seakraken | -57.46% | -10.36% | -10.58% | 12.61% |
| msp-bidbot | -23.78% | -21.42% | 420.23% | -52.58% |
| voterunner | -15.51% | 55.84% | -18.41% | -8.32% |
| mercurybot | 4.10% | 0.91% | 20.75% | -58.45% |
| postpromoter | 40.45% | -9.95% | -4.48% | -34.97% |
| mrswhale | 495.05% | 27.26% | -9.03% | -58.99% |
| smartsteem | -46.61% | 46.18% | 2.08% | -76.30% |
This table is based on samples of ‘last round’ returns on investment taken from https://steembottracker.com/ on 19/01/2018 at different times across the day. This is a very small sample, so don’t judge the bots based on these returns. Use this information as a general idea of the type of returns available at the moment.
So what strategies can be applied to bot usage? I will investigate from the perspective of the content creator (high quality and original content). First a number of questions need to be considered. These questions should include the following:
‘What is my budget?’, should be the first and probably most important question. If you have very little SBD available, then you will initially have very few options. You can buy SBD. If that is not an option, you can try to earn SBD. I would recommend posting quality and original content and then be patient. There are curators around Steemit that will find your content, your content might receive a very high upvote from the Streemian guilds for example. Once you have some SBD, you need to figure what you want to do with it.
As mentioned earlier in the post, the amount of exposure generally depends on what lists your post gets onto. In addition to the lists. Exposure also depends on the number of followers that you have. The more followers you have the greater the exposure your post will get. It is also important to know that it is important to have genuine followers and not followers that follow for the sake of following others. You want followers that will read and appreciate your content. This is why it is important to have the right tags. Even with a small budget, it is possible to get on the ‘hot’ and possibly ‘trending’ lists of less popular, but relevant to your subject matter, tags. I use the tag #economics. It is not one of the most popular tags but it is relevant to my subject matter. See the screenshot below.
From the ‘trending’ posts, it appears I should target an upvote value of around 70 SBD to be able to get close to the top of the economics ‘trending’ list for any length of time. Another tag that I like to use is #philosophy. This tag generally requires a higher upvote value to get higher up the ‘trending’ list. See the screenshot below.
From the ‘trending’ posts, it appears I should target an upvote value of around 100 SBD to be able to get close to the top of the philosophy ‘trending’ list for any length of time. I could still target 70 SBD and get to the top of the tending list for a shorter time. It is possible that the one day near the top of the ‘tending’ list could give you enough additional upvotes to reach the target of 100 SBD and therefore keeping your post up the list long enough to get more upvotes and more followers.
Another option is to change tags. If your post falls off the ‘trending’ list of a particular tag, you could go into your posts and change to less popular but also relevant tag. This should give you a good chance to get higher up the ‘trending’ list of this tag and help you get your post to a wider audience.
Some people post more than 5 times a day, some people post less than once a week, and others post somewhere in between those frequencies. The frequent posters tend to have shorter posts. The less frequent posters are able to author longer posts.
If you are planning on posting very frequently, you have less money available to pay the bots for each post. The best strategy would be to use tags that are relevant but less popular. Your posts will have a much better chance of making the trending lists. You could diversify your subject matter and get on multiple ‘trending’ lists reaching a wide audience at a very little expense per post.
If you are planning on posting infrequently but your posts are long, very detailed, very engaging, and of exceptional quality. The best strategy is likely to involve paying a large amount to the bots and attempt to reach the relevant popular ‘trending’ lists and possibly even the main ‘trending’ list if the upvotes you gain from popular ‘trending’ lists can carry your post that far.
I post around 5 times a week. I had a stretch where I posted every day. I aim to get each of my posts onto their respective ‘trending’ lists. Posts that I believe are particularly important I will pay more to the bots so that these posts can get as much exposure as possible.
Long-term goals are important. I want my content to be as widely distributed as possible while being able to have a steady flow of income that I can use to support myself and my content. So a key goal for me is to attract genuine followers. Consistently getting on the most appropriate ‘trending’ lists is a good way of doing that. The bots serve an important role in getting my content on those lists. The bots can only do so much. I still need to maintain the quality and originality of my content. I also need to read, upvote and comment on other people’s posts on the relevant subject matter. It is important to engage with the people who might become future followers.
Once a good following has been established, the use of bots should become less relevant. The upvotes of your followers should be able to carry your posts onto the ‘trending’ lists. This may take a long time to happen. If the usage of bots continues to grow, the required upvote value to get on the ‘trending’ lists will continue to increase.
Another long-term goal could be to build reputation. A high reputation should encourage more people to follow your content and should increase the number of upvotes that your posts receive. Upvotes from bots increase your reputation.
When should you bid on a bot? As soon as you can. The longer your post has been up, the harder it is to get your post on the ‘hot’ and ‘trending’ lists. Delaying bidding also shortens the time your post has on the lists. If you plan on using bid-based bots, it is a good idea to have a bot or bots that you want to target so that you can secure the upvote from the bot/s as soon as possible. I need a combination of bots that have sufficient ‘maximum profitable bids’ for me to reach my target upvote value as soon as possible. This can be tricky as bots tend to get hit with many bids right before the next vote. There is a good chance you will receive a negative return but it is more important to get on the lists.
There is generally very little point using bots to upvote posts more than 12 hours old. I have, on a few very rare occasions, used bots to upvote posts about a day or two days old. I will might do this if I believe that post is particularly significant and needs just a little boost to stay on a ‘trending’ list.
I have noticed several bots have reduced their maximum post age to 3.5 days. This seems like a very arbitrary number. The ‘maximum age of post’ should reflect the maximum age that an upvote can effectively and consistently promote a post to a ‘trending’ list. Reducing ‘maximum age of post’ would likely reduce profits to bots and increase returns to bidders.
The growth in bots on Steemit is an interesting and quite curious phenomena. It is, in my opinion, the response to the opportunity from profiting from upvote selling. Profit drives activity, Steemit is no different. In the short-run, the growth of bots is not particularly harmful. The bots can help people promote quality original content. Unfortunately, the bid-based bots are not discriminating regards to quality of content. This allows poor quality content to rise on the lists, thus making it more difficult for good content to thrive. The poor quality content will discourage people from joining Steemit and might encourage good content producers to leave if their good content is consistently buried. There are a few bots and bot services that support higher quality content but these are rather few in comparison to the total number of bots.
I am also concerned that the bots will enable those with more financial resources to dominate Steemit. Quality and original content should dominate Steemit not content written by those with the most SBD and Steem Power.
I hope you found this post interesting. This post targets authors that have a genuine desire to share quality and original content. I have explained some strategies that can be used to help make the best use of the bots in combination with several other strategies such as selecting the right tags and targeting particular upvote values.
For more information on game theory, you can access my introduction post using the link below: