What is Tagbot?
Tagbot is a bot for upvoting posts with a specific tag. You can read more about it here.
Currently, for each round, the bot fetches 100 posts from the specific tag (order by created)
and filters them in terms of author rep (ignores authors with low reputation) then upvotes them.
Possible Issues and Proposal
1. The bot may vote the same person many times in a row
- Suggestion Bot collects the list of users who were upvoted on 1st round and they will not be upvoted for the 2nd round. After 2nd round, the list will be updated with new users, so the users upvoted on the first round are now eligible for 3rd round. A parameter for "number of cycles in which a user can be voted again" will be created and can be adjusted depending on the situation.
2. People may use tags that are completely unrelated to their post just to get upvotes
- Suggestion One possible solution for abuse of tags is that every time the bot votes, it comments with something like this.
"Congratulations! This post was upvoted by Tagbot for using the tag ____. Help Tagbot fight abusers by commenting "FLAGGED" below if you think the post is a spam or unrelated to the tag."
- If the bot detects this comment, it will put the user to a temporary blacklist for several rounds.
3. New users (low reputation) who produce quality content may not be rewarded
- Suggestion To support quality content from new users, my proposal is instead of using reputation score, we will collect easily quantifiable parameters to calculate "quality score" for each post. Here's an example.
| Number of words | Partial Score |
|---|---|
| <50 | 3 |
| 51 to 150 | 5 |
| 151 to 300 | 10 |
| 301 to 500 | 15 |
| >500 | 20 |
| Number of Images/Videos | Partial Score |
|---|---|
| 1 | 1 |
| 2 | 2 |
| 3 | 3 |
| >4 | 5 |
| Number of Comments | Partial Score |
|---|---|
| <1 | 3 |
| 2 to 5 | 5 |
| 6 to 10 | 7 |
| > 10 | 10 |
The Quality score is then the sum of three partial scores. As you can observe the score depends on number of words, images, and comments to measure the effort done and engagement garnered by the post. The table above is just an example and can be adjusted to better suit the situation.
This equation can also be used as an alternative
where W= number of words, I = number of images/videos, C=number of comments.
From the 100 randomly selected posts, a quality score is calculated and ranks them from highest to lowest. The bot starts voting from highest to lowest. As the voting power decreases, the one with lower quality post would have lower upvote. A specific function of quality score can also be created to calculate the voting weight. Although this system is not a substitute to manual curation, this is one way to efficiently assess quality.
Benefits
- Same user will not be upvoted many times in a row to give chance to others
- Tag abuse can be minimized
- Quality contents are supported
Posted on Utopian.io - Rewarding Open Source Contributors