It seems like a hopeless task to analyze the psychological state of the developers when they wrote some code that we can’t even see. There might have been more than one conflicting intent anyway, or the intent could have been misguided.
In the case of 30 vs 25 seconds, I could imagine at one point we imposed a max of 25 for live games and never intended to support 30 at all. Maybe there’s some technical limitation, similar to why we can’t support a 30.5s period, and the front-end designer didn’t get the message. But surely whatever the developers say, 30s byoyomi would be a really valuable feature, no?
I concede the point, and would like to rephrase my question.
If byo-yomi periods on OGS ended halfway through the countdown, and users reported this as a bug, would you feel compelled to let them know you disagreed with their semantics?
Many thanks for the link as I didn’t know and expect a so vivid interest in NZ rules. I should encourage for people ignorant like I was to go read a bit the linked threads in that group.
Note that I’m not getting in the mood to switch to these rules as long as iI’ll be afraid of OGS scoring if any jigo results. Not that uncommon on small boards btw.
Link below is a good resume about OGS implementation of the NZ rules
Everyone seems to be in agreement of about the fundamental nature of the issue, but just quibbling about whether it should be called a “bug” or “missing feature”.
I’d be happy to look into implementing this sometime in the next few months (I have a list of things I want to do/fix, and this isn’t first up, but it fits in well enough).
I just left an open question in the bug, which I’ll repeat here.
What should happen in (double) elimination tournaments? A few options:
Black gets eliminated.
White gets eliminated, as now.
Neither is eliminated. Replay the game.
Neither is eliminated. Some other resolution.
Both are eliminated. Someone else gets a bye. (If this is the final match, no one wins the tournament?)
Tournament director chooses up front in settings from a subset of the above options.
Disallow integer komi (NZ rules) in elimination tournaments.
Depending on implementation burden, I might just start off with (2) since that’s the status quo, and having an answer (even if unsatisfying) unblocks handling draws correctly in the rest of the system.
But I’m curious what people think the “right” answer is.
Maybe someone from the NZ group would be able to weigh in? What would happen in a Real Life tournament?
This should be the correct thing to do, and also this is how the elimation tournaments work currently when moderator decides a game as a draw. It looks like the problem isnt how tournaments and rating ystem handles draws per se, but its just that players cannot score the game as a draw, it has to be declared manually by moderator.
I didn’t play in tournaments, but I used to play on OGS using NZ rules. I stopped playing on OGS a few years ago, when it became clear that the scoring bug was not going to be fixed. Since then I’ve played on KGS which has always handled it correctly.
I’d probably come back to playing on OGS, at least some of the time, if it started treating equal scores in NZ games correctly. It can be good to play on different servers from time to time.
What happens now when a game is drawn? Remember that it is possible for a game with 6.5 komi under Japanese rules to end in a draw (e.g. in a triple-ko situation). It is not a peculiarity of NZ rules that a game can be drawn.
Technically speaking, Japanese rules with 6.5 komi cannot end in a draw. A cyclical situation (such as triple ko) can lead to a “no result”, but this is not the same thing as a draw. The game would have to be replayed in order to determine a result.
OGS does not currently handle the “no result” outcome of Japanese rules games either. The only workaround is for a moderator to declare the game as a tie (which it really shouldn’t be) and then annul it (since it should not count as having happened).
I join this. (Game being replayed). Flipping a coin would be the plan B if it’s too much a constraint for the other players of the tournament waiting. Basically a jigo is a success for me so shouldn’t be penalized (well as less as possible)
If there is a game with integer komi that has reached scoring and it is apparent that it should be jigo, currently, a moderator can manually declare the game as “Tie by moderator decision”, which does not annul the game. It would be a ranked draw (if the game was ranked to begin with).
In my last post, I was talking about a different situation, the “no result” outcome under Japanese rules, such as would arise from a triple ko, eternal life, or other long cycle that the players choose not to break from. This “no result” outcome is different from a draw and should be viewed as if that game has not happened. However, OGS does not currently support that outcome, so a workaround is to declare a tie and annul.
I think I would favour “both progress” and if it’s the final game then a reply (i.e. both progress to a further final)
There is occasionally already some dissatisfaction in double elimination tournaments when the winner of losers’ bracket wins the final and the tournament rather than there being a further final between the final two players (who have by now each lost once but not the twice that would otherwise be required to be eliminated)
Focussing on a concrete thing that might be possible especially with @dexonsmith offering to implement…
evidence suggests that Koba is correct: if we implement the possibility of the players agreeing on a draw, not only will players then be able to achieve on their own what currently requires moderator action, but possibly at least some tournaments will handle it.
I remain concerned that there is code like this scattered throughout:
I found this just by searching for “winner”. Goodness knows what the effect should be in this section of random code that assumes there should be a winner, the is just a random sample.
I see. The bug was in the specification, so fixing it after coding is complete is not trivial.
I’m a programmer too, and I absolutely understand the reluctance to make the change.
All I can say is that if you do develop a “fixed” version of OGS, which I assume you would open for beta test at a different URL (maybe test.online-go.com), I’m willing to help you test it by playing several games per day with whatever mix of rulesets you ask me to test.
)
or maybe that belongs in the