Comments on Puzzles

In my opinion, one of the most distinct features which made 101Weiqi explode in popularity is the possibility to post comments on the puzzles, so the author can make it better.

I wish we had this feature to complement the fact that it is incredibly difficult to know what other users’ ratings mean. As a puzzle creator/transcriber, I find it very confusing that sometimes I have near perfectly complete and well made puzzles which get very low ratings, while much more lacking ones receive good ratings. (It would be possible to circumvent this last problem by adding other metric ratings, such as completeness, leveling accuracy, etc.)

Other than server data and load, I don’t think this feature would pose that much of a problem nor complexity of implementation, but, of course, I’m not part of the development team.


If the OGS puzzles section is to compete with other tsumego sites (101weiqi, black to play) there is a lot more work to be done.
Clean up the puzzles section and make it intuitive would help, but puzzle comment would also be nice though.

As you can see this has been requested before.

However I don’t think this has much priority (if any at all).

Totally understand your confusion.


Since tsumegos and problems in general are such a big part of getting better at Go, I wish this had more priority on this server.

And, since OGS is where people also play, a feature like AI Sensei’s generating/transcribing problems from games would be awesome too, creating personalized problems, the highest form of emotional connection probably, easier for learning.


Most players thoroughly dislike tsumego.

Tsumego is important. Once you hit a certain level, tsumego becomes a really good way to improve reading. Perhaps the reason some people dont like tsumego is because the current offering is lacking?

You have to wonder why people choose to use a site such as 101weiqi, which is a pain to use for westerners, instead of ogs or tsumegohero etc… Translate doesnt even work well because text is often embedded in images.The answer is that it is better, despite the language issues.

Go to any western site, OGs included, and you’ll get ‘problem sets’ which consist of loads of problems all very similar in nature. You might get 15 problems in a row all with the nearly identical solutions. This is not enjoyable, nor helpful.

OGS nees this…

  1. One big giant library of problems. Every problem should be given a rank, either by AI, or user consensus. Problems should be checked for duplicates. Problems should be tagged with a category, Life and death, Tesugi, Opening, End game. etc… All problems should appear with a random rotation and random colour. You should be able to mark problems as complete, or favourite etc…

  2. All problems should have a solution tree that can be commented on and added to. Many problems have multiple solutions and sometimes I cannot figure out why something doesnt work. Problems in the current OGS collections do not say why something is wrong, play a wrong move and it simply says “incorrect”. The only sequence of moves they have is the correct sequence, the sequence for why my move is incorrect is needed, perhaps along with some comments.

  3. Users should have a variety of ways for doing puzzles.
    a) Completely Random
    b) Random but increasing in rank, solving increases rank slowly, failing decreases.
    c) Like 101weiqi guam, where you complete 10 timed problems and can then advance to the next rank. Perhaps some kind of leader board.
    d) Random within a range of ranks. I do not want to do 20k problems any more.
    e) By category. Some problems are not life or death, perhaps i just want to do some opening problems.
    f) A choice of time limits for problems.
    g) New problems only.

I genuinely believe that a good tsumego system would attract people to the platform, which in turn would increase the player base.

1 Like

I think it can just be a combination of

  • dull
  • tiring/exhausting
  • not as fun as real game

for example.

I’ve seen that in Wbaduk’s problem sets also. I think that can be a plus for learning/drilling a shape or technique for example.

While I think it’s probably a matter of perspective, it could be interesting to have a large collection that is rated in some way correlated with the rating system, or just a separate puzzle rating.

Possibly not as much as you expect in return for the effort of creating and maintaining such a large set, even if the comments or some kind of user voting system supposedly self maintained it in a way.

That is, even if there’s a big tsumego set here, it probably doesn’t mean people won’t go to fox to play a game?

I like pretty much all of your suggestions. I just don’t think it should be a priority for OGS.

Hi Rye, welcome to OGS. Hope you enjoy your stay here.

You make some excellent points and I totally agree with you that OGS Puzzles needs a thorough update.
I asked our programmer/developer @anoek if it is possible to merge puzzle collections. Unfortunately that is not an option.

I just think that revamping the puzzles section doesn’t have much priority (if any at all) at the moment.

Another issue is that the puzzle section is rather polluted:

  • apart from puzzles it also contains tutorials and joseki collections (think they should be set apart from the puzzles).
  • the quality of the puzzles varies from excellent to absolutely bullshit (there are hundreds completely useless puzzles which should be deleted).

In the last few years I have put a lot of quality puzzle collections into OGS Puzzles.
If you search the OGS puzzle database with my tag you will find them. Enjoy.

If you plan to do something about the OGS puzzle section, there are some others who think it needs an update.
Unfortunately I don’t know the front end from the backend. If so I would have started working on it. But if I can help in some way, let me know.

1 Like

Perhaps a clean slate would be best that runs along side the puzzle collections. Puzzles in the collections can then perhaps be imported manually over time and eventually depreciated. I can see it being a lot of work, especially the duplicate detection.

As for the quality of puzzles, perhaps a kind of ‘flag puzzle for review’ so that the community concensus can remove the bad ones.
Also, perhaps there would be a way to give supporters some added features?

I know OGS doesnt have a team of full time developers working on it, so such a change would be a huge undertaking. Its easy to make suggestions when you dont have to implement them yourself :slight_smile: . I’m a software/embedded developer myself so know all too well.

Are you kidding? At the moment the puzzle’s section contains more than 55.000 puzzles.

If there is a consensus about the possibility of removing lousy puzzles, that would be nice. I am willing to have a go at this cleaning project.
But only if there is a consensus … otherwise it is a waste of time.

By manually, I meant a button press by a user, not manually entering all the moves, that would be nuts. It would allow users to choose which to transfer to the new system.
I would have thought though that it would be relatively easy to write a script that takes all existing puzzles and adds them to a new system, each being converted to the new format. The system would then check if the puzzle is a duplicate of one already in the library.

Anyways, I’ve put in my thoughts. I really do think puzzles are important and should be a priority. I’d help if I could, but I’m not a web dev.

1 Like

Maybe @GreenAsJade can tell us more about this.

Seems feasible, though there’s always devil in the details. It is not “trivially easy” at the moment.

1 Like