Here BebeCadum vs. PuppyGirl in upper half both white and black groups in same region both dead which is impossible.
The territory marked as black, so black groups inside the region cannot be marked as dead.
2 Likes
Honestly, it doesn’t matter. As long as if it calculates the same thing, it will be fine. No need to always complain.
I meant the overall score. It’s not an exploit, so no point.
It is a bug.
It must be traced and fixed. It may influence scoring in other games.
We should not discuss should be bug ignored or not. We could only discuss is it a bug or not.
3 Likes
I tend to agree with this, but assuming that Whites points increase (due to captures) by the same number as Blacks points (due to territory), then the change cancels when taking the difference.
In fact in this example, the score difference is correct.
1 Like
That is true for this particular game. But the intermediate counting is still incorrect.
It happens that an even number of errors leads to the correct result. Errors cancel each other. But it is not always true.
OGS is not limited to Japanese ruleset. In different ruleset score will be incorrect and even could influence the game outcome.
I don’t understand why we need to discuss the importance of the issue. Are you suggesting to ignore it completely?
I’m not trying to imply anything more or less than what I wrote.
I would add though, since the game result only depends on comparing the score difference to 0, this example in particular does not demonstrate that this “bug” can affect the game result.
Is it really an autoscore bug? Or did a player just click on the black stones to mark them dead?
2 Likes
Does it matter? If player click on black stones the area should not be marked as black.
I forked the game for investigation Challange UNRANKED
What I found:
- After 2 passes the game doesn’t go to score phase. The third consecutive pass was required.
- Autoscoring on both ends work correctly
- Just click is not enough to toggle stones state
- Shift + Click causes the issue
My conclusions:
- The cause of the bug is UI on scoring phase
- There is another glitch in pass sequence found
- There was a cheating attempt from one of the players. I do not believe that misclicking may occur with mispressing shift at the same time.
I haven’t looked at the issue in this thread, as I don’t have time to delve into it. However, I will address a couple general philosophical points.
The reason why it is important to take into account the importance of an issue is that none of us has the time to waste on trivialities. In the past week, I have had much less time to vote on reports because of the avalanche of posts in the CM category. Some have been useful, but others haven’t. A better, more efficient strategy is to (1) wait to see whether the alleged bug is causing a substantial problem, and (2) discuss it privately with a trusted fellow CM coder before deciding to bring in the whole body of CMs. For example, the inversion of the log entries for game ending and acceptance/disconnection goes back to the beginning of the thumbnails. Nothing was done about it because it was totally trivial, not a source of confusion. Over the past year and eight months, I have privately alerted @GreenAsJade about many oddities or bugs; some were too trivial to worry about, and many others were easily fixed without a massive discussion.
This assumes that the so-called bug will exist in all rulesets. I see no reason to assume that, especially if its cause is unknown.
1 Like
The issue is not related to CM. It is an ordinary bug. I didn’t report cheater here because it is unclear who is cheating and why.
Yes it does. If it’s players wrongly marking stones, it’s barely a bug.
The area is black because white stones are also marked as dead, so what is displayed is the logical conclusion given the information manually given to the scoring tool.
You could argue that the UI shouldn’t even let players mark black stones dead in a black territory, which would be correct but also really low priority and inconsequential. Players can already reach all sorts of wrong results by manually mis-scoring the game.
4 Likes
Sorry, my mistake. I was under the impression that this post was in the CM category. I think that was caused because this would have been better in the CM category, where unraveling possible score cheating could be discussed more freely.
1 Like
It is still a bug. There is still should be a validation. Accept button should not be active in this case.
And I don’t want to say it explicitly. But it is technically a security issue. Input is not validated at the backend.
Fortunately adversary cannot exploit it for some bad things.
I don’t believe this is a bug. Players are free to agree on which stones are dead. That is how the game works. If they agree the black stones are dead, then they are dead.
It may be a bit absurd to have dead stones in your own territory, but it is possible.
Consider this example:
One may reasonably consider the lone black stone dead (or not). White can certainly capture it. The beauty of Go is that it doesn’t affect the score.
4 Likes
No. Black is alive. If white tries to capture it white loses 1 point.
This is true of all uncaptured dead groups
1 Like