EGF vs AGA pros win-and-continue match

[Kirby]

A loss by time, especially when the one who runs out of time is ahead on the board, is pretty unsatisfying for everybody. A loss by lag, when you actually tried to play the move before running out of time is even worse. I don't know what the official rules are for this tournament, but if lag wasn't included as a condition in the original rules, it's a sticky situation.

The most satisfactory result is if they just finished the game to counting, without any sort of timeout. I guess second best would be if the person timing out were behind when they ran out of time. Then awarding the win to the person ahead on the board seems a little bit more natural, since the person was ahead, and the person that timed out lost by the technicality of time controls.

But here, technically speaking, white won by time. But it doesn't necessarily feel right, since they were basically finished with the game, and black was ahead on points. So maybe we should say that black won.

But that's also not 100% satisfactory, because this sets a precedent for future matches. What happens if a player lags out at move 100? How about move 30? Do we use Leela or something to see who was ahead, and give the win to that player? Should they do a rematch? Or do we give the win to a player if Leela has win percentage greater than X%?

It's a very tricky situation, and I'm glad that I'm not the TD. I hope they had something that explained what to do in case of lag, but I suspect that they might not have - it's somewhat of a weird situation that was probably difficult to foresee.

[jlt]

A possible rule concerning internet lag:

1. Suppose player1 clicks before the end of the countdown, but the server doesn't register it because of internet lag.

1.a. If player2 wants to resign, then player1 wins by resignation.
1.b. If player2 wants to resume the game then both players get (say) 5 periods of 30 seconds byo-yomi. Player 2 gets extra main time if he wasn't yet in byo-yomi.
1.c. Otherwise, player2 wins by time.

2. Resuming the game like in 1.b. can be done at most once during a game.

[mhlepore]

This is unfortunate on so many levels. Both players have compelling arguments to make in their favor:

Eric:
- It wasn't Eric's fault that his opponent timed out.
- His opponent was in his final byo-yomi period, and Eric had every right to keep playing in hopes Mateusz would make a mistake. (to those that have a problem with this, do you also have a problem with "preparing to resign" in the middle game?)

Mateusz:
- Mateusz was ahead, and would have won if things played out normally.
- As far I know, he is very well liked and has never been a part of controversy before this.

There is no universally satisfying solution, but I think a rematch would be a fair compromise.

[Oberlappen]

https://transatlanticgo.org/ no oficial result yet seven hours later, and I see nowhere else something to confirm that Eric resigned.

In the KGS Chat and in the Twitch Chat it where mentioned, that Eric resigned. Maybe someone misunderstood something, but it heared like that.

[Oberlappen]

This is unfortunate on so many levels. Both players have compelling arguments to make in their favor:

Eric:
- It wasn't Eric's fault that his opponent timed out.
- His opponent was in his final byo-yomi period, and Eric had every right to keep playing in hopes Mateusz would make a mistake. (to those that have a problem with this, do you also have a problem with "preparing to resign" in the middle game?)

Mateusz:
- Mateusz was ahead, and would have won if things played out normally.
- As far I know, he is very well liked and has never been a part of controversy before this.

There is no universally satisfying solution, but I think a rematch would be a fair compromise.

Sometimes it´s a decision of estimeate something. A rule could be, that a Jury has for example 3 Options and estimate it, according to Position, Facts and what the Players got to say. I think it´s the best Option, cause, there are things, that will never be really clear.

[Aram]

No need for estimates. Just do what any pro tournament would due if outside unforseen circumstances halt the play.

Wind back the clock, continue play from the position from where the disruption happened.


Both players have a referee/proctor with them to see that they play moves that aren't generated by a bot, so there's already a third party monitoring move input.
Shouldn't be too difficult to say if a outside influence (like lag or disconnect) caused the move not to be played. Specially if the move was entered with 10 seconds left.

[Oberlappen]

No need for estimates. Just do what any pro tournament would due if outside unforseen circumstances halt the play.

Wind back the clock, continue play from the position from where the disruption happened.


Both players have a referee/proctor with them to see that they play moves that aren't generated by a bot, so there's already a third party monitoring move input.
Shouldn't be too difficult to say if a outside influence (like lag or disconnect) caused the move not to be played. Specially if the move was entered with 10 seconds left.

Sorry, but estimates are needed. Maybe you could decide the game is won, maybe not. There has to be checked, if it´s really not the fault of the Player or not. Or to check, how the wishes of the Players are. Maybe Eric says, I don´t want to Play anymore and loses the game.

And it´s especially about estimating, cause you don´t have clear game like that every time. Sometimes you have not much time, you are in Byo Yomi. But it´s a real tactical situation. The time ran out, cause of a lag and it´s result is not ok, but simply start the Situation isn´t ok either.
In time measures maybe you didn´t find a winning move (a difficult Tesuji), but having more time, cause the Situation can´t be fixed immidiatly , gives you the possibility to find the move.

It´s about giving a Jury or a Referee the possibility to estimate. Sometimes your decision is right and it´s fair, but sometimes it´s not. Here is it about something Basic.

[Aram]

A lagging connection between KGS and the player is an outside factor that isn't something that is expected to happen often.

A referee on-site with the player can confirm this, and play can be continued from the position it happened, with the time that was available when it happened.

This is not in any way different from a real world tournament and outside factors. In those situations as well, both parties might get extra thinking time while the referees wind back the clock and restart the game. This is seen as an acceptable and working solution in the real world, why can't it be online? Specially since both have referees with them.

If you loose on time due to not playing on time, you loose on time.. that is how it works both online and in real live tournaments, however, if something outside your control affects the tournament and stops you from being able to make a move, you just return to the earlier position and restart play.

Has nothing to do with if the game is close or not, in early stages or not. If the game is stopped by outside forces, you just return back to the game when you can, from the position where the disruption happened.

[Oberlappen]

A lagging connection between KGS and the player is an outside factor that isn't something that is expected to happen often.

A referee on-site with the player can confirm this, and play can be continued from the position it happened, with the time that was available when it happened.

This is not in any way different from a real world tournament and outside factors. In those situations as well, both parties might get extra thinking time while the referees wind back the clock and restart the game. This is seen as an acceptable and working solution in the real world, why can't it be online? Specially since both have referees with them.

If you loose on time due to not playing on time, you loose on time.. that is how it works both online and in real live tournaments, however, if something outside your control affects the tournament and stops you from being able to make a move, you just return to the earlier position and restart play.

The Thing that it doesn´t happen often, makes it necessary, that it´s estimted. Especially we have even more time, than in a tournament it´s fixed (and to honest in a tournament on a real board it wouldn´t happen that way, cause your move is on the board, there is only about, if the clock is working or not. I answer from a position I got, cause I am a apprentice in the public Service and I know some of the problems happening and I know solutions based on estimation.

Sometimes it´s unfavourable to have just one solution, believe, I know this for sure. It´s different having 2 minutes addited to the clock and maybe 3 minutes to clear the circumstances, than having ours before the problems are fixed. Both can have difficult position and time too. One has a winning Tesuji, that is hard to see. Give him ours he see it, or worse, he already talked to someone, don´t that is Happening something, cause is not fast to fix and than he hear the move that wins (that wouldn`t happen at a normal board).
This makes it definitely total different from your example and shows why an estimation is needed. In an estimation your way can be used, but also other variations, like a total rematch, according to the one, who decides.

[atarihuana]

we still do not know what the proctor said?

[Uberdude]

Sad to see my fortune telling powers were correct:

Uberdude #2 17.04.2019 11:54
Mateusz, please don't play so close to the end of byo-yomi period! You give the kibitzers heart attacks with our fear you will be a victim of internet lag.

I hope the rules of the tournament were explicit about effects of lag in internet play, rather than them making it up after the event. If they are, as is often case, "lag is responsibility of the player" then Eric winning is following the tournament rules. However, those rules do give the AGA players an advantage as they are nearer the KGS server in the USA. I believe Mateusz was playing this game in Kazan, Russia on a not-particularly-good connection. Given that I think his decision to go into his final byo-yomi period and play moves even 10 seconds from the end was unwise from a meta-gaming perspective.

I understand what you're saying, but i disagree about the 10 seconds being too close.
They are playing with a 1 minute byo. In case a player can not comfortably play at the 10 second mark, then that is a rather huge handicap to the player.
It is not something the player should have to think about, specially since this is one of the biggest "money tournaments" in the EGF & AGA.

In case a player can not comfortably play within the rules (play at even 1 second left), then perhaps byo-yomi time control is the wrong choise for the tournament.

Yes, I agree it's a handicap if one player has to play faster to compensate for their bad internet conditions and that's unfair. The question is, do you pragmatically accept that unfairness and adapt to maximize your chances of winning (I expect that if Mateusz kept 1 spare period as "lag insurance" and played fewer moves in the 5xth second then the deterioration in his play would have less effect than the decreased time-out chance), or idealistically continue as though those conditions don't exist and have an increased chance of losing? Considering things 'outside the board' is something pros need to do even in real-life events. For example in his commented games book Lee Sedol says that he had trouble eating oily and spicy Chinese food and was often playing his games in the Chinese league hungry, which he partly blamed for his bad results in the early years. So he took pot noodles and also adapted to eat the cuisine and his results improved.

A problem with not giving Eric the time win (and I say this as a slow player who has many painful losses on time, including in the 3 hours main and 10 stones in 10 minutes British title match) is that Andy could then turn around and say "If I knew the time limits weren't rigid then I wouldn't have played that awful time-suji blunder in my game in a winning position and I would have won". That argument does have some holes (he wasn't winning at the time of his blunder according to bots, it was a middlegame position with lots of game still to play rather than late endgame with a clear big lead) but does have some merit too.

[atarihuana]

i think they should first clarify what caused the timeout. i think it would help to adress the general problem with such situations apart from this specific occaision.

[yakcyll]

i think they should first clarify what caused the timeout. i think it would help to adress the general problem with such situations apart from this specific occaision.

This is the most difficult part of the whole ordeal - since it was a transient error in network transmission, there's no proof of what happened other than Mateusz's and the proctor's statements. Unless the proctor was paying direct attention to Mateusz's actions at the moment of the network timeout, they can't honestly say that Mateusz did submit an action in time. If he also has no directly verifiable proof of doing so, for example through a camera recording of himself playing, it will be near impossible to prove objectively that this was the case.

[Javaness2]

The situation would be easy to manage in an over the board event. If you have a faulty clock, then you ought to be able to stop the game and have it replaced. What are the rules here though? It seems inconceivable that they didn't consider what to do if one player is disconnected. The options are unfortunately a bit limited on the internet.

You could continue through a Skype-type application if you were near the end of the event.
That would mean that the Proctor would have to use a clock to count down 60 seconds byoyomi each move, whilst the two players used a demo board. Ugly, but not impractical.

You could also just restore the position with a 5 period byoyomi and let the players restart on SD.

Pity they didn't have the funds after all to send all 10 players on vacation in Iceland.

By the way - why isn't the Canadian Go Association a co-host?

[atarihuana]

i think they should first clarify what caused the timeout. i think it would help to adress the general problem with such situations apart from this specific occaision.

This is the most difficult part of the whole ordeal - since it was a transient error in network transmission, there's no proof of what happened other than Mateusz's and the proctor's statements. Unless the proctor was paying direct attention to Mateusz's actions at the moment of the network timeout, they can't honestly say that Mateusz did submit an action in time. If he also has no directly verifiable proof of doing so, for example through a camera recording of himself playing, it will be near impossible to prove objectively that this was the case.

no that makes it easy. either you trust the proctor or you do not. (there could also be the possibilty that the proctor did not catch it) assuming they trust the proctor (or why else would you have them?) this specific case should be easily solved.

ps. i am aware that this is not a solution for the general case, since it does open up loopholes. thats why i said its important to distinguish between the general problem and this specific case.

i am glad this happend actually. i hope aga and egf get together and make the best go server possible