I’ve been participating in tournaments for about a month. I have witnessed 2-3 players cheating every week. and it was very obvious and overt. sometimes we feel it, but we can’t prove it. I want to give an imaginary example. performs 10kyu above a player level and the current antivirus system does not detect ai usage. because these players have played some moves themselves or they prefer random moves in between to avoid getting caught. I think this is a serious problem. Also another issue is that the penalty for ai usage detection is just a warning for the first time. (despite the size of the infringing gap I mentioned in the previous thread). Anyway, maybe there are limited solutions to this issue under current conditions, I don’t know, so I don’t want my complaint to be understood as offensive. but I believe that maybe some penalty increases or more effective solutions should be produced. because I feel this neglect is very common.
By the way, I’m giving the number 2 3 only for those who are proven, that is, very, very obvious cheaters.I wanted to underline it again.
sometimes people can play 5-6 kyu better than their current level. or performances may fluctuate due to reasons such as fatigue. I’ve come to the conclusion that ogs’ current policy on this is tolerant. but it looks like it’s being abused.
What you’re describing could be sandbagging (ai-assisted or not). Sandbagging is not tolerated here, and you should report games where you suspect sandbagging to the moderators.
Yes, I’m reporting what I’ve noticed. but some of them are not caught even though they are very obvious. Consider that the user is using 3 kyu bots for example. the current system cannot detect it. I mean imagine a player at a much lower level. Or it’s using a different advanced AI. it cannot be detected by the system. I have already mentioned various crucial examples above.
We do what we can. At this time some of us (including me personally) pay a lot of attention to this. Unfortunately, the botters are many more than I imagined. But it’s like this everywhere
I see this and I am grateful for it. But maybe people can share their ideas about what can be done here and we can reach a synthesis. Which I think will be useful for admins as well as users.
The most that can be done is this: if you suspect someone of botting, file a report with the mods (i.e. for mods’ eyes only) naming whom you suspect and citing some example games of suspected botting.
It is ridiculous that you try to guide me or reduce the options to 1 by yourself. I know what I need to do and what I’m talking about already. Please try to focus on what I’m saying. I emphasize once again that what I have been talking about from the beginning is a serious problem.(!) Also, I did not try to criticize anyone. so I guess this is not the answer I deserve or need. I don’t know how to express myself more clearly.
an account was cheating. but after reaching 9dan this was noticed and banned. yes all sites have this issue and probably ogs are actively dealing with reports about it. however, many users still continue to cheat and a solution cannot be produced because it cannot be detected with the current method. maybe there are suggestions from software experts or people who have knowledge of the subject in the general public sphere. We are on an interactive platform and why not take advantage of it? Isn’t it worth giving a chance?
I don’t know you from Adam. Most people, especially those who are new here, don’t know about the report system, so I was specifying what can be done. It was kindly meant, and you are not the first or the last person who might report a botter. Other people may need the information I gave.
oh wow I didn’t thought there are cheaters even in kyu range, I thought it only happen in high rank
Already said, but I confirm that it also happens among kyu players, even among ddks. Sometimes associated with sandbagging (intentionally resigning/timing out won games/too early), sometimes it is only in some games.
I also repeat the invitation to report any suspicions 
I will add and emphasise that this is topic is ongoing in moderator discussions continuously.
We want to be rid of it as much as anyone else.
I don’t think that we are “tolerant” but we take pains to be reasonable and diligent.
Your “detection method” is your “feel”. Imagine if someone else gets this feeling about you. And imagine if OGS responded with an immediate ban. You would be rightly outraged.
So moderators have to take the considerable time and effort to actually make a concrete appraisal. Assessing moves, comparing to bots and using other cunning tools. It’s hard work, and the likely reward is always dissatisfaction - either the reporter (if the result is not a ban) or the accused (if the result is a warning or a ban).
The approach to complaining about this in the top post is misguided because it does not take this into account.
The top post is written as if the rest of the community is unaware and the moderators uncaring.
If anyone genuinely wants to help brainstorm better ways that we might all respond to AI cheating and sandbagging, my advise would be to take a more respectful approach, in the knowledge that you are not the first person to think of this, and that we are already doing all the things that many clever people combined can think of.
Well thought through and researched input is welcome.
It’s also worth mentioning that for obvious reasons, not every tool used for detecting and confirming AI is is widely discussed. Any discussion of a tool to detect AI use is an invitation for some “curious” person to see if they can get around it, or for hardcore cheaters to take care not to be detectable.
But they exist and are used. This is important to know, because if a moderator tells you that there is not good evidence of botting, then you can be sure that this is based on much more than just “their feel vs yours”. It’s “your feel” vs a lot of time effort and tools applied. Guess who is more likely to be correct…
It’s indeed one of the most draining tasks we have because:
- It’s really hard to feel certain about a decision to ban a player for botting, since there is an entirely grey spectrum between users who don’t bot, user who do it a little and users who copy every move. None of the detection tools are solid, since there are known cases of users being suspicious according to one tool, but not using bots with almost complete certainty. It’s always a sum of several tools lining up that can lead to certainty of botting.
- It’s disheartening, because there are just too many botters among high dan ranks. We have some automated tools that help us detect botters, but frequently when running these tools over the games by one suspected botter in high dan ranks, multiple of the opponents appear to be botting as well. Pursuing those players leads to even more botters. You can spend hours without end going from player to player, giving warning after warning, ban after ban, and it barely seems to help.
- Only about 10% of the botters will actually admit to botting on first being caught. Many will simply deny it, some being adamant about not botting for a long time, but finally admitting after being confronted with extensive and (for them) unexpected evidence. So, basically, you constantly get replies that make you doubt you actually discovered a botter.
- It’s actually time-intensive. When I inspect a user for botting, I often spend 15 to 20 minutes going through games to find enough evidence to make sure that I’m not banning an innocent user.
What about letting botters botting?
I mean: if they don’t do sandbagging, they’ll eventually become 9+dan and will be forced to play against each other.
I understand this could be a problem for actual 9d: they’ll be surrounded and defeated by a bot army.
But from my 4k perspective, botters can do their thing and be satisfied with that (silly asses!)
I’d even extend the dan range up to say 12 or 15d so that botters would simply fly up above any real player’s head.
Is all this effort actually necessary?
Could focusing on sandbaggers be more profitable?
Or am I just dreaming? 
Some players use bots in 50% of their games. Like their real level is 5k, are rated 1d, win against 1d when they use a bot and lose against 1d when they don’t.
I totally understand why you criticize my first post. However, on the contrary, I did not want to go into some sad details and the purpose of my message, on the contrary, is empathy, not criticism. Because 10 15 people were banned in a month because I reported them. Although I reported at least 15 people, it was not banned because there was not enough evidence. I also thought of reporting at least 15 more people. While doing this, I was asking a manager friend of mine. I realized that this job is very difficult and not worth the effort spent. The most basic part that I criticize is that some really abuse it. Who knows how many times he has created an account and continues to cheat. And it only gets 1 warning. I am aware of the importance of user experience, we cannot ban every person we suspect. But I was already trying to explain that they directly attacked this weakness of the current system. Imagine a player using ai against an admin. Without hesitation. This is a big and serious joke. Anyway, the other point I criticized was the insolubility. Glad you’re back with nice details and feedback. In order for you to understand me and my feelings, I ask you to imagine how many people I meet every day. Sorry for causing misunderstanding. Best regards and love…
Also yes, discussing and detailing the current method here may invite people to look for the vulnerability of the system. But if there are 10% 20% players who already use at least 1 move ai in their games, I think we will not lose anything by opening this issue. Thanks for your detail on the root of the problem.