Only activating the Antistalling feature if passing doesn’t change the expected score by more than 0.49 would account for 90+% of this, right?
1 Like
In theory yes. Some tests would be needed to check whether OGS’ AI score estimator does the job, since it’s not always accurate.
2 Likes
What about cases where the staller plays “ko threats” after dame has been filled?
Those will run out; and you have to respond to those anyway
Those will run out
Right, but this is stalling, and the whole point of the button is that you get to forgo the stalling. Keep in mind that these are a bit more annoying than ko threats - since the staller has 3 moves to change the situation, not one.
you have to respond to those anyway
I mean if I’m ahead by 50 and my opponent puts 2 stones in atari, I’m probably not going to respond
1 Like
Those guidelines are not a draft in the usual meaning of the word, nor are the other three guideline documents on Git-Hub. They are in wiki format so they can be updated or corrected as needed, just like all of the OGS documentation. Those guidelines were copied and pasted from the “current” versions in the Community Moderation (CM) Forums category to make them available to the whole community. They represent the current state of the guidelines after substantial discussion, updating, and correction, in actual “draft” versions and in other discussion threads, which can be seen in the CM category.
I have proofed the four Git-Hub guidelines, and found a number of formatting mistakes and typos, as well as two items that need updating (one important, the other minor). I have sent an itemized list to @GreenAsJade and am awaiting approval before making the corrections.
4 Likes
So we make it take a few more preparation moves to activate the Antistalling feature in exchange for preventing abuse of said Antistalling feature? Sounds 100% worth it to me
Whoa, thanks for the prod, I’ve responded.
I want to second this point: these are not “Draft” (in the sense of “we’re working on them before they are official and in use”).
As per "Community Moderation": it's a thing now. those Guidelines are what we’re using. They are “the approved Guidelines”.
The changes I’ve just reviewed with Conrad are teeny typographical matters.
3 Likes
This is a more complicated one, because certainly there can be threats to play that shift the score by huge margins. But I think if the server is looking at the game score, the availability of huge threats does not impact what it thinks the likely end score of the game is because there are simple answers to those moves. If you and I are playing, and you are 99% to win, and I make “ko threat” kind of move that threatens a huge group, the server does not switch to me being 99% to win until it sees if you’re going to respond to it. The server knows that you are going to respond appropriately and the game will not change.
So I think that the “server thinks no move changes the score more than .49 points” rule works fine.
1 Like
OK - so it feels like we have a decent handle on some ways to improve the system.
What’s the process for getting that in front of the people who can actually change it?
1 Like
I mean it’s in front of people that can change it (not me), in this thread.
However we keep conveniently ignoring the counterpoints that I don’t think were unreasonable.
The counterpoint to this is
1 Like
I dont believe your proposed change would be a net improvement. It just fixes the rare case that annoys you.
Are there any stats on relative amounts of “stalling” after game end (no valuable moves left) vs in middle of a game? Either from human moderator reports before this 3-pass feature was added, or of when it triggers now? My assumption was after game end stalling was more common.
I would welcome stats also 
This seems fairly plausible. Since 9x9 games are short, it wouldn’t surprise me if a large amount of stalling in 9x9 games for example were after the game ends.
I’d be curious about the relative frequency of the two for 19x19.
Edit: Since it already depends on board size anyway
it wouldn’t be unreasonable in some cases to have the solution depend on boardsize.
I also think this 3 pass feature triggering once incorrectly is a much worse negative than it triggering once correctly: it is much more rude for a go server to stop you playing legitimate go than a human player to play stalling moves, see Server ending game by its own decision - #39 by Uberdude. So if 40% of triggered cases are mid game, and even if of those only 20% are false positives where the player forced to lose was not stalling but just losing, by my value judgement ratio it would be a net improvement to implement the “only trigger if no valuable moves left” change.
4 Likes
Might depend on how you evaluate it. In principle, I know what you mean.
However, we can always try to adjust how it functions as we’re doing it now, and we can always warn players that find ways to abuse it, rather than use it legitimately.
On the other hand, there’s not much we can do to help players that are being trolled by their opponent except designing another replacement system, or being omnipotent to help at all times of the day. There are ideas for those sure, pausing games for adjucation say by moderators or maybe community moderation, so then both players can leave the game and move on.
It could also be argued, that if we could just change results after the fact, like turning a loss into a win, rather than only being able to annul it, even more flexible systems could be designed, or handling certain kinds of trolling might be easier to correct (results and ratings wise).
So you could be right, that maybe this is just too low of a ratio to be worth helping or prioritising, compared to harming users wanting to play on to the logical end of a game, but I wouldn’t want to rule out helping players that are being trolled in the middle of ranked games.
Maybe some other system needs to change to compensate.
1 Like
I don’t think there is any system anywhere ever that will prevent all annoying things from happening. My goal for changes would be:
- Don’t create new avenues for cheating (which the change to the server decision did do)
- Try to prevent cheating that results in a change in the final result of the game (which I’m talking about)
- Try to prevent other annoying behavior
With what you’re talking about, stalling in the middle of the game, it’s super annoying, but also very easy to counter by continuing to play point-gaining moves. Unlikely to result in any change in the result. I have had players do this in absolute time, and they lose every time.
And on that note I will say – any of this stalling happens in fewer than 1% of games I’ve played. And in those, I deal. I have never made a complaint about stalling. It was not a problem that I would have ever asked anyone to address. And that’s why I find it very suboptimal that a “fix” for stalling was implemented that has created a mile-wide opening for people to cheat in the games that I prefer to play. Anyone, any time now, if they are ahead on points – even only by 10 points – but fatally behind on time, destined to lose, can win their game using a server decision.
3 Likes
My solution is to turn off the server decision on absolute timed games.
What is the net negative in your mind with that solution?
Who, specifically, makes the decision?
^^
1 Like