[ Add-on Support Thread ] 🏆️Anki Leaderboard by Shige

This thread is a support thread for the Anki Leaderboard, for basic usage please read the Wiki.

image02

6 Likes

I think clans are interesting too, but the challenge is that there are not enough users to join groups and countries.

Groups only have about 10-20 users except for medicine and languages. About half of them join countries (so about 1300 users are pirates). So even if we create a clan, most users may not join. (maybe possible in the future if more users join?)

Public groups seem like a good, but private groups may not like or be interested in competing.

So the idea I’m thinking of at the moment is to enhance the national leaderboards as NameLessGo says. It might be interesting to count the ranks of users’ leagues like medals: e.g. UnitedStates: alpha(2) Beta(15) Gamma(30)

Maybe add a built-in recommendation to join clans :question:

1 Like

Could we have pirate players appear in the Global leaderboard :pleading_face:

It might move country players to pirate “faction”, but it would be funny

2 Likes

Suggestion 1:

Suggestion 2:

Perhaps incorporate the “Next Day starts at” for each user in some way. For some it may start at 4 AM. For others earlier or later. It would be help a more uniform comparison

1 Like

It is interesting idea, however, I think this would partially disturb the idea of people comparing against each other in their countries or against other countries.

1 Like

A function similar to this has already been implemented in Country.
If you access the Country leaderboard while you are a pirate, you will see a button to register your country, once you have registered, it will not be displayed. But only about 50% of users are registered in a country.
So if we were to introduce a clans, there may need to be a mechanism to force people to join.

I think the challenge with this is that pirates will always be on top. Currently most users in the US are about 300 users, but pirates have about 1300 users, so once pirates are introduced, other countries cannot win no matter how hard they try.

At the moment when users submit their scores they are calculated and submitted based on their “Next Day starts at”.

All countries vs. pirates might be fine, it would be about 50-50. But my idea is to auto assign the default country according to the language used by the user, so in the future there may be a lot less pirates (It’s still just an idea)

3 Likes

Suggestion: Adding Score column to the main page and make it be default sort order (just like score in the league page).

1 Like

I too would like to add scores but I think the concern with the score is that power users could completely dominate the top. e.g. There are users who have been using Anki for more than 10 years and if their Streaks are included in the score calculation they may always come out on top. This problem would not occur in the case of reviews.

There is already an add-on to display Anki’s data in Discord, NameLessGO gave me the info and I forked it (since the author discontinued). It might be possible if I customize this to display the leaderboard data, or embed a function similar to this in the leaderboard.


add-on: :iphone:AnkiCord - Discord Rich Presence

2 Likes

Suggestion: Add competitive bots to the leaderboard.

1 Like

Huh, Why? Expand your suggestion

1 Like

DerIshmaelite is currently at the top of the alpha league (top league) so it seems there is no competitors. (Maybe posting it in a subreddit for medical students or language learning would increase the number of eager learners.)

I think of developing a leaderboard that works offline, it would be useful when the server is not working due to maintenance, or when the user’s device can’t communicate with the internet, or when the user doesn’t really like to compete online. It would be interesting to have a bot with famous people like Einstein, Mercury, etc.

1 Like

It would be really helpful for the feel of competitiveness to have a review counter that shows up in the menu bar as you are reviewing while also showing the review count of the nearest “players” which updates every once in a while.

image

2 Likes

I already developed a prototype of such functions but do not plan to release it for free yet, because of the increased server workload (probably crashes the server). In short technically it is already possible but server cost is the bottleneck.

So I plan to develop workarounds like these for now:

  1. Make it work offline.
  2. Separate servers.
  3. Allow users to use their own servers.
  4. Auto disable when server workload is high.
  5. Optimize code to reduce workload.
1 Like

@Shigeyuki I have seen this issue brought up now and then. People complain that language learners hoard reviews quickly, so the comparison is unfair to someone studying a more challenging subject. Perhaps adding an indicator next to the name could help identify who is learning what :question: (Law, Medicine, Language, Biology, Geography…etc.)

1 Like

I receive this complaint often but I think medical students actually dominate the top ranking for now. Language learners review quickly but not many users study for long periods of time every day. Medical students take longer to review but almost all of them study longer so their final score is higher.

So far the most common complaint is that some users review too quickly (e.g. 1 to 2 sec/per card or less than 1 sec), so I plan to develop a feature to hide them all at once depending on the number of seconds. (or auto exclude them from the calculation)

As you said I think it would be useful to have a feature like the Flair in Reddit. Or display group names, from the group names I assume that most of the users are learning medicine or languages.


Perhaps the fairest way to compete is to calculate by specific deck. e.g. if a user is studying both language and medicine, the language study group is calculated by the language deck, and the medical student group is calculated by the medicine deck.

1 Like

So far the most common complaint is that some users review too quickly (e.g. 1 to 2 sec/per card or less than 1 sec), so I plan to develop a feature to hide them all at once depending on the number of seconds. (or auto-exclude them from the calculation)

I am that one user, but it is not that I spam that button. I just answer the cards quickly.

3 Likes

Yes, IMO 1-2 seconds is normal and should not be a problem, it is possible if the user uses auto advance like speed focus mode, or filter deck, etc., there are various cases where the seconds can be faster.

In my opinion if it is less than one second it is a bit too fast, but this number of sec depends on the users preference, there are sometimes users 10 to 20 seconds/per card.
So I think it is reasonable to allow users to set their own seconds.
e.g. a user who is reviewing slowly can compete with users who are learning at roughly the same pace by hiding all users who are reviewing at less than 10 sec. In this case other users are just hidden so it does not affect their learning.

1 Like

@Shigeyuki I believe this might be a bug? I studied more then 5m/day however my streak is only 24 days.

1 Like

Hello thanks for reporting the problem.
If you customized “Next day starts at” you will need to set it manually. Please check if these settings match:

  • Leadearboard: Config → Settings → General → The next day starts […] hours past midnight
  • Anki: Tools → Preferences → Review → Scheduler → Next day starts at […]
1 Like