It is actually serving to flush out some ambiguity in our definitions.
It is very clear that “Joseki” in OJE have to be Credible, with everything that we’ve discussed applying.
It is not so clear (in the documentation) whether this applies to every move of a sequence and it’s marking as “Ideal”.
It is not explicit whether every Ideal move has to be Credibly Sourced.
What it does say is:
“This […] cannot be shown as inferior to another choice.”
My opinion is that we would be best served if we tighten this up to mean “This cannot be Credibly shown as inferior”.
Or in other words, anything marked “Ideal” must be part of a Credible Joseki, which surely means it has been played by pros, or worst case documented by them even if we can’t find where they played it.
Note: the converse is NOT TRUE. Just because something IS a part of a Credible Joseki does not mean that it is Ideal. A move may be part of a Credible Joseki, but demonstrably inferior in the general case to some other Joseki move. This would make it Good not Ideal.
Edit: I’ve linked this post into the OJE documentation as an official supplement on the policy.
I don’t think you need to search on an empty board, but for reliable results in Waltheri’s, you should use a bigger search window like this (or perhaps even one line bigger) to filter out middle game situations and opening situations where there are already stones on the side:
When a variation is already not joseki, because a parent move was dubious, how are moves marked in a perfect continuation?
For example, in the continuation from https://online-go.com/joseki/18862 all moves are ideal and I cannot see that I’m already outside “joseki land”. I have to navigate back in the variation to see that black is already tricked.
I don’t think it is a new realization. That is pretty much how it is - I’m not aware of many variations that are entirely ideal that don’t lead to Joseki:Done. On the other hand, there are continuations after the joseki that are ideal which might be caught by a bulk update like that.
Edit:
There are actually quite a few moves that kind of branch out too quickly to put in established sequences that are ideal but not leading to joseki. Like there are 10ish games with a joseki move but then there are a handful of responses that get 2 or 3 games. There are also branches that end in a “See Position:6109” that would should maybe also have that end node tagged as Joseki.
Well - one thing that is fantastic for sure is that the person who has contributed by far the most to OJE already believes that moves that do not lead to Joseki should not be Ideal.
This means that the bulk of our content will already be correct in that respect - yay!
FWIW it is a new realisation for me. I had not fully appreciated that it would be a incorrect to put “Ideal” on the best move that follows a mistake in OJE.
EDIT:
I think that what you are observing is that these moves have been added in the knowledge that they are Joseki, and are marked Ideal for that reason, even though they don’t currently lead to a Josek iposition.
I have to assume/hope that these do lead to a sourced node, even if that node is not the end of a Joseki sequence?
Otherwise they actually don’t belong because they aren’t sourced and therefore aren’t Credible…
Oh, I don’t understand this. If there is a Good/Trick/Mistake move then I think you can still have Ideal responses. I guess they could all be changed but that makes things difficult. Say you have a ladder variation that might be labeled Good because it is situational - I think you will also want the responder to have ideal and good moves too.
Yes, sourced by pro games predominantly. This joseki is an example. Up until the end points, it is played enough by professionals to be credible but then it branches out enough that the next move is not established. So it doesn’t get a Joseki:Done tag. I guess the issue is that Joseki:Done has some connotation that it is finished but that’s definitely not always the case.
If the realization is that every full string of ideal moves is joseki then that is my thinking too. Doesn’t have to match up with the joseki tag, I guess.
And I haven’t always added the Traditional/Post AlphaGo tag because the idea is that the branch will get more established and you didn’t want sources on intermediate branches.
I think it’s because it’s up to interpretation. If I wanted to see the follow up to a mistake by black and I saw the only response was a ‘good’ one for White, I’d be think wait is there a better answer not included?
The other issue is that, more than likely if we can conclude that blacks move was a mistake, there’s a good chance black playing any of the follow up moves in the sequence are also a mistake. Black would either tenuki or play a damage control sequence instead. So you either get a sequence of mistake, good/ideal, mistake, good/ideal etc or it’s all good or all ideal.
I mean you could say white is ideally following the punishment of the mistake, or Black is ideally following the path of the mistake.
But anyway once there’s a consistent policy it’s good.
