Escaping by Timeout

You can already withdraw, though it will count the forfeits as losses.

My suggestion there would be to simply introduce unranked tournaments. All the fun without the hassle.

1 Like

Here is the way I think: problem → anlysis → solution
The rule is the last part. Problems where here before the rule: cheaters, escapers, timesettings, ranking, rudeness and so on.
If the solution isn’t good, let’s change it. But let’s keep clear what comes first.

I would also add something that doesn’t apply specificly to GaJ: I am tired of this discussion because it seems to me that 80% of this topic is about PRIDE.

He’s an escaper: he offended me!
He’s a timeouter: he offended me!
He messed up a tournament: he offended me!
He doesn’t believe me: he is offending me!
He doesn’t remind what I said before: he is offending me!

There are countless statements in this topic that only refer to pride, with colourful expression like “is a slap on my face” and so on…

So many people feeling offended, annoyed, betrayed, tricked, what else?
I’m here to play Go and enjoy the game. My pride is harmed enough by my real life. I don’t want to waste my time worrying about a loser whose only satisfactions are to inflate his rank and to escape when he’s losing a game. Poor man.

I sympathize with players that have a busy life and still try to play in the remnants of time.
I sympathize with those that didn’t understand that a “weekly” tournament can take over a year to finish.
With those that didn’t figure in advance that multiple rounds can start unpredictably and submerge you with tens of games.
With those that didn’t understand well the difference between resigning and just leave.
I also sympathize with those little cheaters that can’t find better joy than trick the rating system to feel stronger.
And I sympathize too with mods and TD who have to deal with all those cases.

If some math is working badly, let’s try to fix it. But it’s impossible to fix everybody’s hypersensitivity.

6 Likes

Exactly and people don’t want to do that so they time out instead (someone mentioned further up). My thought was that if there was a third way to leave tournaments people wouldn’t need to time out. And then any remaining time outers would be disappearing people, who we presumably are not much concerned with, and cheaters, who would then be more easily identifiable. I guess there will always be a few temporary life event cases but at least the pool of suspects would be reduced.

1 Like

Xalexander is “exhibit A” in my opinion. I have looked at all the games in all the clusters in the most recent 150 games. The large majority are losses, some are inferior positions, and only 2 or 3 are wins. The wins appear to have been sacrificed so as not to break the string in each case. All this, plus the repetitive pattern over a long period of time, is decisive evidence of cheating, I think.

1 Like

About resigning games in tourneys: 5d to 6kyu!

Wouldn’t it be better to timeout and annul the streak of games?
Will that have consequencies on global rating?

An apology, some thoughts, a few ideas and a workaround.

To all those who have persevered with this thread, let me start by saying, ‘I am sorry for those occasions when my words have gotten harsh.’ I feel confident that everyone involved shares a desire for OGS to be the best it can be.

There’s a few ideas that I would like to stick together in this post so grab a cuppa :coffee: or skip the post if you prefer.

It seems to me that the challenge we face is creating a system where players with very different approaches to the game can coexist productively/enjoyably and we want to achieve that without disrupting the ranking system’s ability to accurately assess Go-strength. Some players want to treat every game like a titanic struggle for the fate of the world (ie. seriously) while others wish to ‘take it easy’, ‘go with the flow’ after all ‘it’s just a game’ and not stress when they have other priorities (ie. casually). Neither of these extremes is inherently wrong but they work better when paired with like-minded players and of course, there is a full range of players in between.

Idea A

Incorporate into the profile page some kind of indication as to the attitude of the user.
This site: g0tStats? already provides something of the sort including:

but unfortunately that site doesn’t distinguish between blitz TO’s and correspondence TO’s which are of course very different.

All I’m talking about is a number or rating of some kind that reflects the likelihood of a person timing-out of a correspondence match. No diagrams would be required. It could be as simple as looking at what % of their correspondence games over the last n correspondence games have been lost by TO. That number could then be normalised for the general OGS population, or not.

Just on it’s own, this idea could assist mods in automatically identifying the most extreme examples without waiting for complaints to come in and for that, the stat doesn’t necessarily have to be visible to the general population. If it was visible then players would have the option of looking at it before investing time and effort into a game.

Idea B

For tournament games, when player A’s TO% is significantly lower than player B’s. Give player A the option to change the game to unranked or cancel it before they play their first move. A pop-up something like this when they first open the game:

Your opponent has a significantly higher tendency to time-out than you. Would you like to:

  • A: Proceed with the game normally
  • B: Change the game to unranked
  • C: Cancel the game
    Nb. game results will still have the usual effect on the tournament but will not effect your site-ranking in the case of B & C.
    [Box] Tick this box to never receive this option again

and a tick box for it in settings.

It’s just a crude mock-up but you get the idea. Maths wise all we are doing is removing a little bit of data some of which would have been garbage anyway so negligible effect on the ranking system.

The main advantage of this idea is that it empowers the first player to decide how seriously to take the game. They can choose to unrank it (B) and then have a casual game with their casual opponent or even if they proceed normally with the game (A) they will feel like it was their choice.

I don’t really see this idea taking off because I’m guessing there would be a significant amount of coding involved but maybe someone else can evolve the idea.

Idea C Workaround

Allow players to cancel tournament games before playing their first move. Currently; if you are White and Black has already moved, you can’t cancel, you can only resign with full consequences (in tournament games). I have sought clarification of this in this thread. Dare I say it, the current situation actually distorts the ranking system.

Soon after I joined the forums I was made to realise that for many, tournaments are just a convenient way to arrange new games in bulk. The tournament result may be utterly irrelevant. We already have Quick-Match and Custom Game players who prefer to check out their opponent first and then cancel if they don’t like what they see.

So i’m asking the mods who read this thread; As a workaround. If a person resigns a tournament game without playing their first move, is it reasonable to use the in-game call mod button to request an annulment in those cases where it is not automatic?

5 Likes

I like your idea A a lot. With regard to the question in idea C, we have done several requested annulments recently. Also recently a few canceled games failed to automatically annul for some reason and have had to be manually annulled. So I see no problem with requesting an annulment in such cases.

2 Likes

The player who motivated this thread now timed out after I stopped preventing it with my own holiday time:

Take note that this is what you are defending here: Cancelling a clearly lost game at move 235 against an opponent of weaker rank and alternating the time-outs of lost games with a string of wins.

As far as the recently proposed solutions go, this game, like most of my games, is not a tournament game. I send out a challenge and someone takes it. I certainly do not see why solutions are proposed for tournament games in particular. This is not a tournament problem. It is a problem of allowing cancellation of correspondence games. I am also not against displaying the time-out information, but I do not see how I would use it after the opponent has taken my challenge.

1 Like

The solution could be so simple…

It should really be normal that a MOD can change an outcome of a game. (I do not understand why this still isn’t possible.)

Then people could report games like this. It may take some time for the mod to arrive, yeah, but the Mod could declare the timeout as loss. Problem solved.

Of course, Mods shouldn’t do this on a whim. But games like this, which are clearly lost will be counted then.
And i think it would be REALLY effective against this sort of abuse

1 Like

Mods can decide a winner if the game is ongoing and annul a finished game. They can’t retroactively make an annulled game ranked.

Since it’s quite time-consuming to investigate someone’s history and it’s more likely an accusation is false than true, I doubt “let mods decide” is a good approach.

Also, you might say “the game is clearly lost” but then where do you draw the line?

I know they can’t but i still do not understand why. It’s completely senseless that a mod can’t do that.

They don’t have to.
Seriously, why should they invest a history? They should just look after a game which got reported and change the outcome. This may even be done in under 1 minute. The benefits would be worth it.

Just look at the game and tell me my assumption is uncorrect. What’s wrong with changing the outcome of missing data to correct data in the system?
I personally also do not really care where the Mods draw the line, the main problem is it should be possible first.

1 Like

I’m saying that just because there’s 1 game for which you can tell the outcome, that doesn’t mean it’s a good idea to make it standard practice. Players already have a hard time telling if something’s seki or bent four.

If someone plans to time out with ill intent (and this is the only relevant case), they will just time out before the result becomes obvious. It wouldn’t solve the problem and just put a higher burden on mods.

p.s.: Why is the game history relevant? 1) To know in the first place whether a timeout would result in annulment and 2) to make an informed judgement call on whether it’s deliberate or not. If it’s not about to get annulled, there’s no issue with letting the game time out. If it’s not intentional, the player might come back to finish the game or they will also time out of other games, in which case we’re back to the issue of protecting rating integrity.

It is a good standard practice to get help from a Mod if something is wrong with the game. Also, there is no seki, there is no bent4, there is a game which is already completely finished. You can try to argue to the moon and back, that game is definitely timeout abuse and it should be possible to handle it by a mod.

This is wrong on so many levels. Why should someone time out when they don’t even know they are losing?
And what burden? There aren’t even many abusers at all, so how is that a big burden to take a quick look?

and i say it again: the history is not relevant.
There is a game which is already marked as time out. It doesn’t matter if it’s deliberate or not. The player can report the Timeout, the mod takes a quick look and changes the result - that’s it. If the result is unclear, the mod doesn’t have to do anything.
This also is a betterment of rating integrity, because lost games will matter.

1 Like

Clearly losing situation and not massive timeout and I don’t get rating gain. I don’t mind massive timeout that are simply quitting the site but deliberate timeout should be treated as losses. If I don’t see the site working on this I will just quit this site.

1 Like

How do you know that your rating wasn’t changed? I have no idea how to see the effect of a single game on a rating.

Hello @lamma8888, there is little point in blackmail :wink: reasons for and against the rule are listed somewhere above. If you have an idea on how to solve them, you would make a lot of people happy. But I am afraid some compromise might be inevitable.

Yes, I think most of us would agree with that, but detecting those cases automatically is rather challenging. Which is why for now we are sticking to manually reviewing reported cases (thanks for the report by the way.)

Does this mean that all correspondence timeouts are unranked, for the time being?

If that is the case, I do think there’s something very seriously wrong and exploitable with that system.

However, it was my understanding that timeouts counted normally up to some small number N of consecutive timeouts (where N was somewhere between 3 and 10, but I don’t remember).

So, how does this currently work, exactly?

First timeout counts normally, following consecutive timeouts are annulled.
please read at least post 136 before answering (as not to keep going in circles) where I tried to summarize the thread so far (feel free to read more of course, but I assume the thread is now unreadably long for any normal person :slight_smile: )

Indeed

I have indeed been following this thread lately, and read some of the responses.

A little bit of context (and disclaimers): I’ve been playing on OGS for a long time, since before the merge (does anyone even remember what that is? or call it the merge with a the?). I’ve carried the T of shame on more than one occasion—hell! I might even have contributed to coining that term—. Now seniority means nothing, it’s just that I’ve seen a multitude of rating systems, with different rules, being implemented (where a multitude is greater or equal than 3).

Back in the day there was no safety net to break your fall. The system was extremely punitive, not only did it drop you to the bottom of the ocean (there were lots of people lower than 30k, I myself lost 10 to 12 stones once), but it also put right in front of your profile “This player has lost N games on timeout (X%)”. It was harsh. The effect: I sandbagged… a lot! So did many others.

OGS ranks were probably the hardest among servers (probably harder even than East Asian servers). There was a strong suspicion that the policy on timeouts was pulling all ranks down. Also many players spent pointless months crushing low ranked players. There was such a thing as a “strong DDK”. In short, ranks were all over the place and, in general, dragged downwards.

Now, before making the assumption that all of this was “deserved”, consider that, often times what happened was that I had a bad couple of days, and when I came back I had an almost unusable account. After all, timing out in bulk is pretty much expected, more often than isolated time outs.


So these are my reasons for not wanting to go back to such a system. However, I understand that the problem here is intentional cheaters. People looking to artificially inflate their rank, while denying legitimate rating points to their opponents.

In my view, ideally, we wouldn’t care about ranks. We would only care about being reasonably paired for matches. After all the rating system is not a Tournament, or a Ladder, or a King of the Hill; it’s just a measure. But we don’t live in such idealism, for better or worse we care about our precious rating points, we (many of us) feel rewarded when our rank increases after a win, and punished when it decreases after a loss. Most of all, betrayed when we don’t get an expected reward, that’s even more frustrating than losing.

With that in mind, I also don’t want a system that, by these context, “rewards” a cheater, thereby encouraging them to continue their behavior. I also wouldn’t particularly want to “punish” them, because that is not the point. In my opinion, they should more-or-less break even in the end, rendering their whole effort pointless.

For that reason, I think it is important to distinguish the behavior of a cheater from that of a bona fide player that runs out of time. So, after all this verbosity, I do have a few ideas to propose.

This ones I consider at least reasonable:

  1. I do think that rating only 1 consecutive timeout is awfully low. Perhaps it’d help to increase that number to 3 or 4. I guess that might drop you a couple of stones (just a hunch, I have no evidence for this), but it’s not a tragedy.

  2. Admins should have the power to flag a player and having all of their timeouts be ranked from that point onwards (as long as the game itself is ranked). That way, for all accusations, the case is reviewed and offenders wouldn’t be able to cheat anymore. Much like bans, this could be reversed if appropriate.

  3. A properly ranked player would have about 50/50 wins and losses (again, no evidence, but this ratio, or any other, could possibly be determined with a large database, such as OGS’s). So, when a player times out, you could poll their last N ranked games (for some reasonable N, for example N=20), and see if their win ratio is higher than 50%. If so, the game should be ranked. The reasoning behind this is that such player could have been stacking their losses for timing them out in bulk (somehow).
    Alternatively, instead of polling by games, it could be polled by time (for example, 3 months, or what I call the length of your average correspondence game).
    This alternative has its natural limit when the player equalizes wins and losses, which (combined with #1) might be about right.

  4. Automatic vacation would be awesome regardless. More of a discouragement for cheaters, who would actually be losing something, and great for those moments when life just happens but you want to continue playing. May I ask for this to be optional in the settings, or would that defeat the purpose?

And now, the crazier ones:

  1. Before the first timeout is set as unranked, this would have to be approved by referee (subsequent consecutive timeouts are set as unranked from that point). This could work if A) Timeouts are rare and B) There aren’t many cheaters around.

  2. There’s a sixth idea, but it involves putting a strong bot as referee to determine if it’s a clear loss or not. I don’t think that’s workable for the time being.

So, @AdamR , I’d like to know if you’d consider backing up any of these.

Cheers

EDIT: I do not actually know how big the problem is. I’ve not encountered it, but I’ve come back from a long hiatus, so it may just be that I haven’t played enough games lately. I do support the old adage: “if it ain’t broke, don’t fix it”.

8 Likes