Ingo wrote

One question, in the light of the possibility that
9x9 in the next computer olympiad has some chance
to become a tournament with integral komi (= 7.0 ).
Is there a way to let at least some of the 9x9-KGS
bot tournaments be played with integral komi?

Thank you. That is a good idea. I can run the April and July 9x9 tournaments with 7.0 komi, as practice for Bordeaux*.

Petr wrote

 One thing to consider, can the KGS tournament system even handle jigo
results? Won't we be hitting strange bugs etc.?

We might hit strange bugs, it will be interesting to find out.

 Hmm, thinking of it, maybe Nick even mentioned once that he cannot
change the komi settings in tournament setup... but I'm not sure of
that.

I can set the komi to 7.0. That is, I can set it to 7.0 in the tournament definition file which I upload to KGS. Whether the KGS tournament system will respond correctly, I don't know. Finding out, in advance of Bordeaux, would be a good idea. Maybe not just with the official KGS April and July tournaments, but an informal test tournament, using some of Aloril's pathological bots, and fast time limits, just to see what happens.

Here are three things which I already know the KGS tournament system handles wrongly - or at least, not the way I would like.

1.
I cannot set a tournament to more than 20 rounds. This is deliberate, If I try to set it to 24 rounds, it explicitly refuses, and tells me that 20 is the limit. I would like to run KGS 9x9 bot tournaments with more than 20 rounds.

2.
If I specify certain overtime settings, it uses different overtime settings, without any warning. For example if I specify Canadian overtime of 10 moves in 30 seconds, it uses 25 moves in 30 seconds.

3.
If two players are paired for the second time in a tournament, their colours are again chosen randomly. They ought to be the reverse of the previous pairing.

All of these have been reported, more than once, to William Shubert ('wms'), who is the only person able to fix them. He has not fixed them, but has said that he can fix 3.

He often rejects proposed changes to KGS (e.g. Fischer time) because they require changes to the user interface, and this will involve communicating with users. I sympathise with this; but none of the items listed above will involve any changes to the user interface, so I am hoping that he will agree to fix them.

Olivier Teytaud will be running the bot events in Bordeaux. We have agreed that the 19x19 and 13x13 tournaments will be played on KGS, using the KGS tournament scheduler; I will assist him with this, though I will not actually be in Bordeaux. Olivier does not plan to use KGS for the 9x9, because of 3. I am hoping to see 3 fixed, so that KGS can be used for all three Bordeaux tournaments.

Best wishes,
Nick

* Bordeaux is the city in western France where the European Go Congress will be held this summer. See http://www.egc2011.eu/index.php/en
--
Nick Wedd    [email protected]
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go

Reply via email to