I was thinking about a way to gamify (somehow) MusicBrainz editing.
The idea is quite simple: each action as a MB editor has a number of experience points associated with it.
Basically, adding a release could add 100 XP, adding front cover art to it, 10 XP, adding a note to any edit 5 XP, adding a link 10 XP, having an edit voted No -50 XP, commenting on another editor note 10 XP, etc… exact numbers have to be determined, and could be adjusted (actions are already stored, values for each could be a new table, and adjustements would be possible).
It would give an XP score for each editor.
Status could be based on this (ie. beginner 0-5k XP).
New ranks could be created.
It would instantly give an idea about how much an editor is experienced, with much better granularity than current status.
And the score could be used to “upgrade” an editor to auto-editor, ie. when reaching a certain score a vote will be automatically created.
Yes, but i don’t think that will happen much, do we have people doing posts on this forum just to get some badges ?
Also we could have an admin priviledge to give a negative XP on bad behavior.
Never heard of D&D Level Loss because you pissed off the dungeon master ?
This already happens and i don’t think XP would change anything in this regard, editors having bad behaviors would loose points at least, and are always at risk to be banned.
While i understand they are always people trying to do shit, the vast majority of editors are respectful to other editors and guidelines.
On the other side, it would encourage people to add data, quality would still be controlled by the vote system and peer reviews.
But for me the main point is that it would allow us to give more privileges based on the score, automatically.
I like the idea and I also don’t think this will lead to more misuse either - unless you make the gamification element so great that people come here just to play the game and not to improve the data. So don’t add dragons and vampires and we should be fine.
I don’t like the idea of negative points for ‘no’ votes. A well-intentioned new editor could end up going negative (or at least staying at zero) which could be discouraging.
My first thought was that the XP wouldn’t make much difference to me, but then I remembered that I do check my edit totals on my profile page fairly often…
Not really a huge fan of the idea. Rewards need to tilt towards Quality over Quantity.
Main con I see there is some people are just driven by scores.
The scoring for comments needs to actually score valid comments not just someone pasting “a comment” into the field.
Maybe the AE’s also get the ability to adjust scores if someone if clearly gaming the scoring system.
Also what about the real experienced editors? Are they starting at zero XP points?
Quality bonuses maybe needed?
Trouble is some artists are so much easier to work on and research than others. Again I can see a burst of people diving into the common artists to fill them with obscure edits instead of quality research.
Or fighting over the game of “first” on new releases.
A big problem I see are the scripts. I can see people hauling huge amounts of data in from odd sources and not checking it fully.
It already gets a bit odd when I am working on something and suddenly I find a mass merge script triggered in the middle of my work. I’d worry that may happen more often if people are chasing for high scores.
And no rewards for bots.
(random thoughts from the head of a tired person after a long day)
That’s a good point, probably not a good idea to reduce experience, but we could have a slow down if you get No votes too often. For example, points earned will be divided by 2 for a week, or something like that.
Well, it isn’t always a problem, quality of their work might be. Some are, not all, by far.
The idea was just to give a small bonus for a note in an edit, we can’t measure “quality” of the note. But the lack of note is more often an issue to me, than “bad” notes.
Yes, the idea was to have all actions having a basic score, than can be adjusted over the time.
Nope, we have edit history for all editors, the actual score would always be calculated from it somehow (we count actions, and apply a weight to them), then cache the total value to avoid extra calculation.
If weights are modified, we proceed to a complete recalculation. ← @bitmap, @yvanzo any technical issue to do that ?
Not sure what you mean here. But i see you are worried about quality, as i do, the goal here is not to give points based on “quality” but more on “activity”, because in practice, more an editor is active, more the quality increases (someone editing since a long time is likely to have a better knowledge about guidelines and various cases, knows how to research and to make useful edit notes, etc…).
To improve quality we need to make edit review easier, guidelines clearer (and more easily found by newcomers), to have more editors (and more voters), to have an undo system, to have better UI, etc… the XP idea don’t address those, it just about adding more granularity between editors, we currently have only 3: beginner, normal, autoeditor. I want something like 20 levels, someone with 500k edits wouldn’t be on the same level as someone withe 600k, someone adding cover art to many releases, but never adding a release would still gain XP, someone adding links but never doing a merge would still progress.
That’s a legit worry, but the score will not give anything by itself, i mean nothing more what we have now.
It would just allow an automatic management of privileges given to editors.
Beginners will still have restrictions, etc… but there will more “levels” betwen “normal” and “autoeditor”, and even more between “autoeditors”.
Currently i’m worried by the fact a lot of experienced editors don’t have a chance to become autoeditor (even though they are doing good work since years) just because they aren’t “known” (and therefore never suggested for a vote).
The XP system could improve that, because we could just ring a bell when an editor reach a certain level, and trigger a vote.
And it might the way to protect vs abusers: to access level N, you need enough Yes votes from editors from level N and more, as we do for autoeditors. (just an idea)
Bots are identified, so easy.
That’s the point of this post, trying to imagine improvements for a beloved database. So random thoughts are welcome
Regarding some sort of penalty for No votes: I think it would be better if the penalty was for edits voted down. That way if an editor cancels his edit, or argues his case such that the voter changes his vote or other editors upvote the edit, there’s no penalty.
Yes, it was what i had in mind when writing about “No” votes, it’s rather edit voted down. Also points cannot be given before the edit is accepted (as it can be cancelled).
I think there should be more weight given to edits that have to pass the voting queue than to auto-edits. Also, what if an edit was voted down but generated a lot of votes both directions and a big discussion?
I’m personally on the fence about this idea, but I have to comment on the following:
While majority of editors would probably qualify for that statement, I have seen quite a few editors with edit numbers ranging up to a couple of thousand edits (mostly add release or add relationship edits) AND who have been editors for years who didn’t even split the featuring artists from titles into the artist credit or didn’t lowercase “The” or “A” so I wouldn’t really call these editors experienced…
Voted down edits: I think it makes sense to “punish” those - but maybe only if the edit got at least 2 no-votes and at least 66% no-votes. This way you likely won’t get punished for little mistakes or edits where the opinions are divided, but rather only for real bad votes. I think it may even be a good idea to have a new rank for editors with negative scores - those would be the editors suspected of spamming or trolling and all their edits should be votable and maybe they should even need at least one yes vote to pass.
An auto editor could be able to set them back to 0XP if it’s clear that they learned.
Auto-edits vs. Open edits: Someone before said autoedits should count less. I’d say auto edits without an edit note should count less. Auto edits with notes should count the same as votable edits, but maybe edits which get accepted with 3 or more yes-votes and no no-vote should get more points?
Edit-Notes and comments: I think it’s a good idea to reward leaving edit notes on your edits, but there should be some kind of system in place to rate the quality of those notes. For example there is one editor who leaves the edit note “by Junkies” on all their edits - that is useless. Others just type in “.” for all edits that you have to leave a note on. Also edit notes like “merge” for merges are not very helpful I think.
Comments on other people’s edits on the other hand don’t usually contain nonsense, so they could get points either way.
Voting: Not entirely selfless, but I think you should get XP points for voting too.
Maybe even for forum activity too?
Versatility: I think it would be great if you could get bonus XP points for being versatile. Maybe double points for the first ten edits of each type and a bonus 500 XP points if you made 10 edits of each type of edit that editors without special privileges can make.
I’m not a fan of gamification. I think it works fine as long as both getting XP and the bonuses for having XP only affect the user. Because then they can only screw themselves.
A good example is Duolingo: the experience is a great driver to get you through the tree and learn a lot. But there is also a section where people can translate articles from the language they learned and review other people’s translations. Cue the hordes of idiots who paste in Google Translate results and have no clue about editing. The only reason for that is XP, because there’s no way on earth it helps you learn anything.
The concept of better ranking is a nice idea. When having conversations with editors it will let newbies know that the person they are talking to is worth listening too.
Working through cover notes and research to add in all the various instruments and relationships can take hours. Especially with proper checks.
Pressing a button to import from Discogs without even looking at the results takes seconds.
This is what I was meaning by “bonus quality points” as not all edits are equal.
I’ve seen some editors go on big complex trawls to clean up artists and various sections using lots of external research.
And then others who will just wade through hundreds of changes of hyphens.
Again work which requires different types of skill. I’m not knocking the hyphenators, still valuable work, just that it takes more hours to do the research.
