Escaping by Timeout

OGS Development
1

Could you please finally stop the “feature” that people can simply time-out their lost matches in bulk?

The argument that it would “distort” rank if time-outs count as time-outs, but it does not distort rank if people time-out their lost matches is ridiculous. Some of them may not even do it on purpose, but simply do not know what move to play … because the game is lost.

And no, I certainly do not want to complain about individuals, because there is nothing intrinsically wrong with timing out, it is not an expected rank behaviour that games do not count after timing out.

7 Likes
2

I have never bothered to get involved in the threads about this before, but it’s happened to me a couple of times in the past. People have bailed in debateable or losing situations when the risk to their rank has been high… I’m not saying every one of them has been a deliberate escaper but the situation is not ideal.

1 Like
3

Is that even still a thing?

OGS needs a changelog.

1 Like
4

For the sake of readers new to this topic, let me add a note of clarification. This thread is referring to a rule that applies only in correspondence games. Timeouts in live games are of two kinds. In legitimate timeouts (often in blitz games), the player runs out of thinking time and times out. This counts as a loss and is recorded as a timeout. A timeout that results from escaping (disconnecting from the game without resigning) counts as a loss and is recorded as a resignation, because it is indeed a dishonorable form of resignation. Escapers who are reported get warnings, and habitual escapers are banned.

4 Likes
5

Sorry, so what do correspondence time outs count as?

6

So you mean that instead getting the “T” after first timeout, all the following timeouts would impact my rank as they were resignations?

It might be fair to the system, it might be fair for all the other players, but for an user it would really suck to come to OGS after few days of no internet and find my rank 5 stones lower than it was previously. Guess that wouldn’t be that big of a deal with this glicko2, since ranks tend to balance themselves out pretty fast.

7

The rule under discussion is that the first of a series of timeouts in correspondence games counts as a loss, but that subsequent timeouts in the series do not affect the ratings at all. Typically, strings of timeouts occur due to a sudden life event. such as, to take an extreme example, dying. Or, in one case I suspect, parents who ban a child from playing.

2 Likes
8

Hello,
well while we know it is an unpopular rule for many and I understand (and always have) that for an individual it may on occasion be plain unfair, I am afraid that AFAIK there is not removal planned for this rule at the moment. And as far as I personally am concerned the rule is largely a good thing (stone me :P)

As far as I understand it is not, and it is only part of the reason behind the rule. Very simply put yes, there is a difference between feeding the system potentially very wrong information or omiting parts of it.

it is common for players with many correspondence games to stop playing (altogether, or just for a while)

if all of the timeouted games of the player counted two things would happen. First, the player’s rank would get completely devastated. Quite possibly hit the floor completely so to say. That is not a big issue, unless the player comes back. Sometimes people just take a break, have some pressing matters or just forget to set vacation. If the player gets back to his original profile, he/she will have to climb all the ranks back, you may say it is his/her problem (largely so), but unfortunately it will impact everyone he/she plays to get the rank back, basically sandbagging for quite a while, ruining other’s games, mood and rank in the process.

Second, especially if he/she was playing against considerably weaker opponents all those would get potentially undesrved rank boost, which is probably not a huge deal, but given enough of these, it might eventually start being a problem. If we annull those games the ranks of the affected players will be slighlty less accurate than they could have been with a proper information, but they will not be wrong.

If we keep the rule, it is usually like one unfairly uncounted game for a player in a long time. The only real problem is that the player may understandbly feel “cheated”, but it is almost sure not to cause any real problems. Even if someone really takes the effort to cheat by bulking losing games, he/she can only prevent a rank loss, but will not inflate their rank by much anyway, because he/she will still be unable to win against stronger opponents.

Therefore I still view it as a better choice to keep it, feel free to try and explain to me where my logic fails, I will be happy to be corrected (although the choice is not mine anyway even if you sway me on your side :wink: )

In the previous thread about this there were some alternatives suggested which I found very interesting and we made sure anoek saw them. He might be considering some of those still, what I am saying is that AFAIK just canceling the current rule altogether will probably not happen and I apologize for it being rather annoying to many of you, but I/we feel it is the better/safer option in most of the cases.

However if you suspect anyone of misusing the rule and really bulking clearly losing games together while finishing those “good” ones, please do let us know. While I think it is not a common way of cheating I will be more than happy to look at each and every such a report.

6 Likes
9

Another common case is to forget to activate the vacation mode before going on vacation.

To recover of such a rating drop, let’s say accidentally time out of 15 games, you would have to win at least the same amount of games before your rank is back to normal. If you now play against weaker opponents (your ranking is lower for match making) it will take even longer.

+1

I think if someone mass timeouts of all losing games, one could easily inflate their rank by 2 to 5 ranks (only guessing here)

Either they are really good at obouscating their doing or I never encountered one of those. Lately I check everyone’s profile if they timeout of one of my games.

10

I was passionately in favour of getting rid of this rule when I first discovered the loophole.

But after the subsequent debate, it emerged that nobody could actually point at an instance where someone had definitely done this to their advantage (IE as a cheat) which made it seem a lot more of a moot point.

I think that interest could be stirred up again if someone could actually find the concrete examples of people cheating this way.

GaJ

1 Like
11

Why is it important to be fair to the player who times out at the cost of being unfair to the player who does not?

1 Like
12

“Typically, strings of timeouts occur due to a sudden life event. such as, to take an extreme example, dying.”

No, it is not “typical” that strings of timeouts occur during a sudden life event like dying.

13

"If we keep the rule, it is usually like one unfairly uncounted game for a player in a long time. "

If it were the case of one uncounted game in a long time, I would not have started this thread. It is not. This is not a case of “logic” but wrong statements told as fact without any basis. (I suspect that your basis is “I don’t remember it happening to me and/or I like to time out and don’t care about the opponents concerned by it.”)

And yes, it is perfectly possibe to inflate one’s rank. And no, it is not a very good proposition that others can cheat me by design and that you “are happy to look at the report” if I put in a lot of research work in every single case (which is NOT once in a long time as you claim). I am sure that it is “better and safer” for you if you simply let cheaters do what they want.

1 Like
14

Not this cheating business again…

1 Like
15

You missed “an extreme example,” a bit of gallows humor, but not an unrealistic concern for me,

16

"and find my rank 5 stones lower than it was previously. "

If we suppose that your loss rate is close to 50% in the given time and you would find yourself 5 stones lower if you actually lost all the games you would have won, this means that in the current system your are taking a total of 5 stones from those opponents that would have won. But that is ok because it is not important what is fair for the others?

17

I believe that an argument about being fair to the player who timed out is spurious. What I mean is that this is not the reason that the rule is in place.

The reason the rule is in place is because most people leave the site within 15 games or so.

And often they do so by timing everything out.

So if this happens regularly, as it does, and it impacted rank, then our rank pool would be inflated by it.

That is the reason for the rule.

1 Like
18

I am sorry, this statement I really do not understand. I feel like you percieve the issue as personal to me for whatever reason, but I can assure you that I have nothing to gain by wanting to keep the rule. I would not be the one coding it or having any work connected with it what-so-ever. If anything having the rule deleted might even be easier for me :smiley:

I only wanted to be transparent with you about the rule probably staying active for some time, and my understanding of the reasons behind it. I am sorry that you do not like it, I really am, but I was just being honest. I think the pros and cons have been discussed enough in the previous thread and do not see much sense behind doing it all over from scratch. Nobody is saying that the rule is ideal, but unfortunately a lot has to be based on compromise in such a big “society”. You seem to be arguing that the other way around would be completely fair and simple for all, which I am sorry simply is not true.

My statement was mostly based on not having a single report for such a thing for as long as I remember. If you feel like you are having such a problem often please then maybe we could agree on an arrangement where you send me a PM or a report anytime you win a correspondence by timeout? It should be almost no work for you and I will look at them, see if the game did count and if it is possible abuse? Then if we see that there really might be a problem it might be better ground to try and change something instead of pointless arguing over potential issues/realities. It will not be immediate solution, but if there is a problem it might be grounds for looking for a better solution in the future. Might that be acceptible compromise?

19

If other people tell me that it is important to be fair to the player who timed out and I respond to it you should not tell me that the argument ist spurious. This is obviously the reason why many players support this rule: because they personally profit from it, not because they leave the site within 15 games.

If what you way were true it would be a very simple thing to restrict the rule to the first 100 games of a player, so I doubt that very much. It is about people who time out a lot and stay and do not want to lose several stones because of it.

20

I’m not saying it’s more important, but i am saying that without the current “T”-thing, the effect would be far greater for the poor player who times out 20 games in a row, than it is for any of his/her opponents.

With current system, if i for example were to time out from 20 games, i guess about 10 players would lose the rating points that they otherwise would have gained from me, while other 10 players will save the points they would have lost by losing to me. This is if we think that every game has about 50/50 change of either player winning.

If OGS were to take away the timeout-safety, then those 20 timeouts would have a huge impact on my rank, and that distortion would start to trickle onward when i’ll start my climb back to my “real” rank.

Note that i am speaking purely about “normal” bunch of timeouts that happens (like if i get hospitalised suddenly,) not about a situation where one player racks all the games he/she is currently losing on short of time. I’ve not yet seen anyone actually doing this, so i’m not sure if it’s a real problem or not.

1 Like