In my case, I passed three times, and the server decided that my opponent won.
I think three passes in a row should be done by the winner, not the loser.
1 Like
Why? What difference would it make?
Despite the marging being so large there, I think you passed too early.
M19 D13 G15 were all worth something there so you should play those before you pass. Doesn’t really matter on that game, but if the score would have been within 2-3 points on your advantage, then you would have lost because passing too early.
1 Like
The reason given by Anoek is that it is safe for the winner to pass three times in a row and the winning rate is still greater than 99%.
However, the current situation is that the loser’s three consecutive passes will trigger a server decision.
This should be inconsistent with the original design of the system, and I think the mechanism of the loser passing three times to cause the opponent to win feels very strange.
If the winner thinks the game can still continue, then why not?
5 Likes
Also without moves in those areas, how would you count @BattleOh ? Is M19 a point for black?
It’s normal for beginners not to see such points. Once you do see them you should either play on until everything can be counted properly or - if you don’t want to play those moves and have no hope of winning - resign.
2 Likes
I’m a beginner.
I actually thought the game was over and didn’t realize I was supposed to move it.
3 Likes
That’s fine and entirely normal. And I can see your point about this feature being triggered here is not how it was discussed.
TBH I have mixed feelings about this feature and I have actually less headaches about it in this scenario. The passing player is claiming they don’t need to add any moves anymore, so I don’t see many dangers.
anoek probably phrased it that way because that is the normal way in which stalling occurs. You didn’t answer the crucial question: What difference would it make? The answer is none.
You were 200+ points behind, and your opponent could have passed three times and still won. Why did your opponent keep playing? Some people will do that to give you the opportunity to resign, rather than have the size of your loss displayed. That is meant kindly. As they whittle away your position, they hope you will get the hint and resign. Alternatively, some opponents keep playing because they like to turn the knife when crushing someone. In all cases, the functioning of the new feature puts an end to the proceedings.
Yeah, its easy to miss those kinds of small end-game moves, but those are important in order to count the correct score at the end.
D13 was probably the most important move on that board, if black plays it they create 1 point of territory at E13, but if white plays D13 you can capture 1 stone and and put 3 other stones in atari, so black will probably answer D13 by playing E13 and then you would have so-called “half-point-ko” on C13/D13
M19 would also capture one black stone, and if black doesnt react you could also destroy the small territory around N18/O19. So if you play M19, black would most likely answer on N19 to keep the small territory intact, and then you would have another “half-point-ko” around L19/M19
G15 is just a half-point-ko, there isnt any territory depending on that but whoever takes the ko and then fill it will gain 1 more capture than the other player. The reason why those are called “half-point” kos is that it requires 2 stones to get 1 point as the capture, so 0.5 points per move.
If instead passing you play the sequence wD13-bE13-wM19-bN19, there would be 3 half-point ko’s on the board. Only once all those 3 half-point ko’s are filled the score can be counted correctly.
2 Likes
I hope you understand that this is a new feature of OGS, implemented to avoid endless games when someone keep playing even if he has no chance to win.
That’s not the case in your game and it results in ending the game when you would have liked to see a proper scoring (i guess).
I can understand your opponent choice too when the message appears.
In fact, the winner can cancel the message and continue. His choice.
From the system, that’s a new incentive to not teach endgame to beginners in cases like yours.
Luckily we have the forum to complete the missing moves (thx at @_KoBa)
2 Likes
Indeed. Quote from game:
[21:35] jiggly cow[6k]: ?
[21:36] jiggly cow[6k]: what happened
I think the messaging could be improved. Start with an intro like "Is your opponent being annoying and pointlessly prolonging the game? If so you can force it to end now. " and also have a cancel button “Cancel, I want to keep playing”. Current dialog is unclear of the purpose of this feature and that you can decline.
6 Likes
Recently my opponent and had that message when I was simply closing all remaining Dame points while he (many many stones stronger than me, and obviously winning) resigned three times in succession … that didn’t feel good … especially as I have learned that it was “good style” to close all neutral points before passing.
1 Like
Hello !
I think this was mentioned in a recent thread or post, but I can’t find it.
After pausing a live game, the ‘You are leaving a live game…’ message still appears, and attempting to open other games (ex. correspondence games) in progress redirects to the paused live game.
It seems that this could be problematic if both players have agreed to pause the game for later, as it would be impossible make moves in correspondence and other games in progress with the forced redirect, even with an agreed pause in a live game which could be continued days or weeks later.
7 Likes
I had not seen this change, but just had it come up in a game. Game was scorable, and I passed, but the opponent made a few more moves to kill dead groups. People do that from time to time, not sure why. He wasn’t stalling, as he was winning the game. After three times of me passing, and two moves of opponent killing groups, I got the “Accept predicted winner” button. Here’s the thing, though – I didn’t press it. I was confused and looking around for what the other options were (I didn’t see a way to dismiss it) and boom – it gave him the server win.
How did it happen without me agreeing? Doesn’t it require that I agree to the scoring, if it is trying to protect me to someone stalling?
I admire the effort to curb abuse, but I’m not sure this implementation is the answer. I play a lot of people (relatively) who are tired of a game, either because they’re winning or losing by a good amount, and they just start passing when the game is nowhere near closed. This could reward them for that.
Additionally, I play mostly absolute timed games, which are certainly the most most readily abused on time. And if I’m playing someone who is ahead on points, but critically behind on time – will not be able to finish the game – it seems like they could just start passing, have the server score the game by labeling me a staller for continuing to make moves, and then win the game, escaping a loss by time out.
Is that correct? Or am I reading the situation incorrectly?
Regardless, thank you truly for your committed effort to bettering the server. It’s an inspiration.
6 Likes
I think your opponent might have been the one who clicked the button. I don’t know the details myself, but some of the posts earlier in this thread from people reporting on their experience with this feature seem to imply that the button to unilaterally force game end appears for both players, not just to the player who passed 3 times.
If this is true, then it could be considered slightly odd, at least in the most “prototypical” pattern of stalling. In the prototypical pattern, there would be one player repeatedly passing and trying to properly end the game while the other player is taking a lot of time playing moves that don’t accomplish anything. So if this feature is to exist at all, it does make the most sense to offer it to the player who is passing repeatedly and presumably trying to reach the game end.
But if it does also gets offered to the player who is playing repeatedly, it’s a little interesting. Since in the prototypical case, the player who was passing and wanted to finish and score the game does get the ended game, but from their perspective unexpectedly cut short without the chance for them to properly score when they were actually willing to end the game and score.
5 Likes
Having only seen one side of it, that was my impression of what happened.
And again, looking at points only, there doesn’t seem to to have any big negative impact to have the server deliver the 99% resolution of the game. But with a timed game, it actually creates more avenues of abuse than it solves.
2 Likes
Is there any avenue for abuse?
Let’s look at the situation from a utilitarian perspective.
When the choice to push the button comes up (3 passes, 99% winrate), either both players are unhappy, only one is unhappy, or both are happy to remain in the game.
If both are unhappy, it doesn’t matter whether they push the button or not. But if they do, it can only be a net positive, as the worst that can happen is that both remain unhappy.
If both are happy, the button doesn’t matter because there is no incentive for either player to push it.
If only one player is unhappy, they may push the button and become happy. This may cause the opponent to become unhappy. However, it is impossible to make everyone happy, and therefore this is the best result that we can aim at. From OGS perspective, all players are equal. If one of two must be unhappy, it doesn’t matter which one.
You see, the feature is impossible to abuse.
Edit: my argument has some glaring problems, can you find them all? 
3 Likes
Additionally, I believe it is not much easier to abuse the rule in a timed game. Determining whether you are ahead by 3 passes + 10 pts isn’t feasible in a fast game unless one player is comfortably ahead.
I do think the system should be updated so that the button only shows up to a player who has passed repeatedly (not when the opponent passes repeatedly). The onus should be on the player who is ahead to figure out whether a pass will be detrimental.
2 Likes
My concern is about the clearcut determined at 99% by AI. (+10pts)
Not sure that every case at 99% is as clear for humans as for an AI.
Well that’s a recommendation from some go servers to make sure the autoscoring will be right. Because the score won’t change in some Chinese kind of rules by adding stones while the other pass.
I see that’s troublesome with the new feature.
You may lose the opportunity to score the game when you made it to its end (and wait for it).
A well thought anti stalling feature should never interfere in a not stalling game.
4 Likes
This is in this same thread (see the OP )