I can agree with this maybe in the future there will be easier tools.
I think single separate game should not be considered as âsandbagging gameâ
sandbagger is the one who doing it systematically or his rank will not be affected anyway and he wouldnât be able to become sandbagger
I remember few games where opponent created too complicated situation, so I was confused and resigned or time-outed while thinking where to place a stone. And I consider these games as clear win of my opponent. Because he managed to create situation which breaks my mind - even though kata bot thinks that I was ahead.
7 Likes
That case isnât really within the scope of this topic, but since itâs already being discussed, there were 25+ recent games involved in that case.
2 Likes
When a mod gets called to a game, I can understand that they would anull the game in some cases when something fishy is going on.
But I donât really understand the argument about protecting the rating system. If âundeservedâ wins and losses are really such a threat to the rating system, wouldnât every game result need to be verified somehow, instead of only the small proportion where a mod gets called?
12 Likes
I would argue that the answer to this is ideally âyesâ, but practically we donât have the ability to parse all games at the moment. Even many actually malicious sandbaggers manage to slip under the radar for pretty extensive periods, so this would need some kind of automation (which is a serious plan, automated sandbagger / airbagger detection).
However, when one canât solve a problem completely, thatâs not a good reason to then ignore it completely, is it? Itâs not a choice between doing nothing or doing everything. Every little bit helps.
Although I do agree that a single game is probably not significant enough to really have any effect.
7 Likes
At the risk of going OT again, this sounds remarkably like:
I wonder if the something that is being done is really the right thing to do.
And so how many games are needed before it gets significant?
a long and boring ramble
And what is really the problem?
Is sandbagging the problem or the symptom? Is attempting to keep the ranking system correct in fact doing it more harm than good? (Leaving aside the unintended consequences in terms of customer experience)
It seems that the problem is in fact unfinished games. Sandbagging is finishing games prematurely, escaping is stopping playing prematurely (I donât know how airbagging worksâŚ)
So maybe itâs better to try and tackle that more?
And it seems that quite a bit of annulment is only potential or inadvertent sandbagging.
So my proposal there would be not to bother annulling games unless they are clearly sandbagging and even then annul in bulk at the point of banning that player. Now I realise that there are time limits on the annulment action so this is maybe not something that can be done yet but still that is not a reason to dismiss it.
Meanwhile, we could see what happens if fewer games were annulled. Maybe just issue âwords of adviceâ and leave the games be. Hopefully the saving of mod time would be a side benefit too.
Or have some auto annul thing like bulk time out that applies to games ended before, idk, 100 moves. Ideally with the possibility to reinstate annulled games for particularly short games.
I realise we get a new category of troll -the 99 move quitter- but again maybe some words of advice to those few is worth the time saving for the others?
Finally, maybe itâs just that there are not as many sandbaggers as it seems. If most of the breaches are etiquette ones or committed by children or others who just donât know any better then let it go?
Sorry for that, bit tired now and not able to structure or conclude my thoughts. But trying to consider all alternatives for the sake of completeness. Iâll stop now!
4 Likes
To split hairs, sandbagging is artificially decreasing rank.
4 Likes
I read most of the posts, and I have to say that when I first saw the topic title early in the game when it was made I was like âoh, no, the poor fellow lost all that time waitingâ. By now the amount of time invested in this topic has really blown out of proportion, even though things seem to be about matters that have small significance to the vast majority of the server which is definitely not sandbagging.
Since we do not have all the data and most of us do not know what really goes on in the daily schedule of a mod and what kind of things and behaviours they have learned to look out for, maybe we should cut the moderators some slack and offer them some more trust in their decisions. 
They are volunteers after all and I, for one, think that people that offer some of their own free time for the community and the health of the service provided by the server in general, deserve quite a lot more than just cutting them some slack.
Just my two cents.
10 Likes
I mean, every game is significant, itâs just that a single game is not going to break the system. But Iâd say that if a user is throwing more than three or four games within a short period of time, that would be reason to ask them to change that behaviour (and annul the games), before it becomes habit.
Not really, these concepts are unrelated. Sandbagging is any strategy that will result in a player being ranked lower than they are supposed to, for example by escaping from games prematurely, resigning games where you have not evidently lost, or (the worst kind, which I consider malicious by default) accepting many games to quickly resign them / playing a couple of horrible moves to bring yourself to a losing position.
Escaping is just the practice of letting your timer run out.
This is what happens when complete beginners accidentally win their first couple of games, which results in them getting matched to SDKs or even Dan players, who then resign out of frustration, keeping the player ranked high. However, the term is more commonly applied on what they have in other servers where people can choose their starting rank.
All three of these things are major annoyances for many players, since we receive many complaints and reports about them. Thus any ranked game that comes to our attention and that contributes to skewed ratings should in my opinion be annulled, if not to directly change the effect it has on rating, then at least to show the moderators are doing their best to counter sand- and airbagging.
This is not how our policy is, though. We warn first, sometimes even twice, before banning. Most users are not aware of doing anything wrong, and we generally give the benefit of the doubt in these cases. Even banned users can appeal, and if they react understandingly, they are given a second chance.
However, if a player is playing as a 15k, while being a 10k when looking at playing strength, the games should be annulled there and then to bring the player back to their normal level, otherwise the problem remains: itâs a player who appears to have the wrong rank, and gets matched to opponents who donât expect the player to be that strong.
For the outside, this would just look like weâre not doing anything to combat sandbaggers. I think this idea is silly.
This doesnât work, since there are legitimate reasons for resigning a game before 100 moves, e.g. when you completely screw up a vital L/D problem, misremember a joseki, etc. Number of moves is only partially good at finding out which games are resigned legitimately and which ones arenât.
Sandbagging is a very commonly heard complaint, though: people get very upset about playing sandbaggers. Granted, many reports are not actually provably malicious sandbaggers.
4 Likes
I donât think you two are disagreeing here. @teapoweredrobot is suggesting that sandbag games only be annulled in bulk. The only time one could bring a player up from 15k to 10k would be to annul several games.
@teapoweredrobotâs qualm seems to be with the annulment of a single game with the justification that it somehow maintains the integrity of the ranking system (maybe it does maybe it doesnât)
5 Likes
I have only skimmed this thread but I donât think this has been mentioned and I hope it is a productive suggestion for @Interestinggame:
It may well be worth your while to familiarise yourself with one of the more flexible clock settings. I myself use fischer for everything simply because it reduces time wasted. A good game can be played with 3m + 20s upto 5m which means:
- Each clock starts with 3 minutes. Your clock will run when it is your turn but you will have 20 seconds added to it whenever you make a move. The clocks have a maximum of 5 minutes on them at any one time.
- If your opponent vanishes at any stage of the game you will only have to wait a few minutes for their return or loss.
13 Likes
Thatâs completely misplacing of the issue:
We are not attacking the mods, we are indicating that maybe, just maybe, they can be wrong.
The fact that we mostly like and appreciate them doesnât mean they are immune from being wrong.
Mistakes can happen, and even long lasting and long standing practices can be mistaken (see recent rank adjustment discussion).
And since, as you say, we donât have all the data, we canât know, as you declare, what does or doesnât happen to the vast majority of the server.
If people believe something seems wrong and needs to be discussed, invoking âmod armorâ to shut down the discussion is somewhat in poor taste.
People seem to disagree; interpreting that disagreement as attacking the mods to shut down opposing opinions isnât to the best interest of the community in any case.
Inb4, I really hope we donât get to âif you donât like it then leaveâ, this is the most counterproductive argument.
8 Likes
Haha! The ultimate play - a âbrain breaking tesujiâ, LOL 
4 Likes
Except that Iâm of the opinion that we arenât wrong here. I still donât see the game from the OP being annulled as a mistake, and I simply happen to disagree with the idea that games without a clear and just winner should be ranked. This seems to be a fundamental point on which several of us disagree.
I think as with so many discussions about ranks, the best we could do is explain our position (so far has been done quite sufficiently in this topic, I believe), and then agree to disagree.
I donât believe that is happening here. I think having this discussion is a good and healthy part of OGS, even though you managed to get me quite worked up about this earlier today 
In all cases that I remember where discussions were shut down on the forum, that happened either because it got too heated, or because of external circumstances (OP being a known troll, to prevent public shaming or because the discussion was about personal matters irrelevant to public discussion, such as account deletion or ban appeals).
7 Likes
Iâll ask this for the sake of the argument, not to pump anyoneâs blood pressure 
:
(maybe I am missing something obvious and the comparison will make me understand)
If a clear and just win only should be ranked, then why arenât obvious misclicks forcibly accepted as well or games where they occur annulled?
It is very probable that they are not human error, but a glitch/ bork or involuntary hand movement for a number of reasons.
Shouldnât we also be asking âwho would want to win by misclick?â
If Iâm not mistaken, there are people who maliciously play garbage at some point, to sandbag, same way some lose on time.
If a slight tremor happens and I play A1 when it makes no sense to do so, and this leads me to lose the game (itâs not unheard of for a move to make or break a game) and my 99k opponent refuses to accept my undo and they win, wonât this negativity affect the ranks?
We usually agree to disagree here at some point anyway, because the other option is for all of us to be perpetually moody and nobody wants that. 
8 Likes
You do have a point, there⌠I think there is some argument to be made here regarding accountability, but Iâm too tired right now to put it into coherent words, thus, good night for now!
6 Likes
HA!
I won by timeout.
17 Likes
I think this discussions raises some very fundamental questions:
- What is sandbagging?
- At what point is intervention necessary?
Before this discussion, I was under the impression that it would only be considered sandbagging (and require some sort of intervention) if the behavior (of artificially reducing oneâs rank) was deliberate. However, from this discussion, it seems apparent that the moderators define this more broadly to include (and act on) cases where such behavior may be unintentional and even if the player might be unaware of what they are doing.
I think this broader categorization raises some difficult questions about where to draw the line.
@stone.defender gave a good example above to illustrate this:
@Gia also made a good point recently above about games ruined by misclicks.
Imagine a relatively weak player who is making their best effort, but still they occasionally misjudge (or even just miscount) the position and resign prematurely (when the game is far from decided, or even when KataGo believes they are winning). Should these cases really be considered a form of sandbagging that requires annulment? Or are they merely performing at their level of strength? One could even view resigning prematurely (or erroneously) as an extreme blunder. It would be difficult to tell such players to avoid making such mistakes, when they are genuinely using their best judgement.
By annulling such games, and discarding these instances of their misjudgment, I believe one might even actually be biasing the rating system to essentially airbag them, since, in future games, they are still liable to repeat such mistakes, which were not properly measured by the rating system due to annulments.
13 Likes
Forums are correspondence timed 
5 Likes