Page 1 of 3

Predict the winner of World Go Championship 2017

Posted: Thu Dec 29, 2016 9:29 pm
by pookpooi
Is it too early? Actually, if someone makes poll before Bonobo then it's really too early LOL.
But since we know all players now, let's put instinct to work ;)
I had guessed slightly wrong on Lee Sedol vs. AlphaGo match and very wrong on Cho Chikun vs. DeepZenGo match, hope third time's a charm for me.
*This poll allows re-voting, I don't think anybody will change their vote based on what's other people is voting though.

More discussion is going on in this thread
The original announcement
News on other two professionals

Re: Predict the winner of World Go Championship 2017

Posted: Sun Mar 05, 2017 9:45 pm
by pookpooi
website that has details in English is up

http://www.worldgochampionship.net/english/

Re: Predict the winner of World Go Championship 2017

Posted: Mon Mar 06, 2017 6:27 am
by DrStraw
I am surprised that with all the resources they have and as long as English has been used in go that that page sounds so stilled. Don't they have fluent speakers to write these things?

Re: Predict the winner of World Go Championship 2017

Posted: Mon Mar 06, 2017 8:57 am
by pookpooi
My comment is that I don't like time format.
Absolute time should demolished in go already, it's too strict, and oriented toward computer.

Re: Predict the winner of World Go Championship 2017

Posted: Tue Mar 07, 2017 7:59 am
by dsatkas
yeah, absolute time is stupid for humans

Re: Predict the winner of World Go Championship 2017

Posted: Tue Mar 07, 2017 8:27 am
by Schachus
Is it clear, it's absolute time? "a countdown starts when the remaining time become 5minutes or less" sounds to me like it could be 5*1min byo-yomi at the end.

Re: Predict the winner of World Go Championship 2017

Posted: Tue Mar 07, 2017 8:46 am
by Uberdude
As Schachus says, I expect "countdown" is how they translated byo-yomi. We wouldn't want the humans losing on time because Zen keeps playing nonsense moves inside their territory! The lack of a lunch break is more concerning to me: I hope they are at least allowed to eat/drink at the board even if that means poor manners and a risk of greasy fingers on nice stones because brains don't work well when they are hungry.

Re: Predict the winner of World Go Championship 2017

Posted: Tue Mar 07, 2017 8:46 am
by HermanHiddema
Yep, definitely 5 x 1 minute byoyomi.

Re: Predict the winner of World Go Championship 2017

Posted: Tue Mar 07, 2017 8:57 am
by gowan
I just wonder how time handicaps could be done. What if computers were really short-timed, such as one second per move? Of course the human players might have five minutes per move. Would that equalize the playing field?

Re: Predict the winner of World Go Championship 2017

Posted: Tue Mar 07, 2017 9:27 am
by Uberdude
Time handicaps for computers are almost meaningless unless you also restrict the processing power, because a program that is X strong on Y computer with 5 minutes of time per move could be the same X strong with 1 second a move if you give it computer about (in terms of number of 0s) 5*60 = 300 times as powerful as Y. There are of course economic limits (and eventually "no super-computer that big exists yet") on increasing the power, but you can't plug more brains into Park/Iyama/Mi to make them play as well in 1 second as 1 minute. Also don't forget you'd also have to essentially double the human's time if you want them to have the same as before as they can no longer usefully think on the bots minuscule time (of course you have to turn off pondering for this).

Re: Predict the winner of World Go Championship 2017

Posted: Tue Mar 07, 2017 9:45 am
by Schachus
Uberdude wrote:Time handicaps for computers are almost meaningless unless you also restrict the processing power, because a program that is X strong on Y computer with 5 minutes of time per move could be the same X strong with 1 second a move if you give it computer about (in terms of number of 0s) 5*60 = 300 times as powerful as Y. There are of course economic limits (and eventually "no super-computer that big exists yet") on increasing the power, but you can't plug more brains into Park/Iyama/Mi to make them play as well in 1 second as 1 minute. Also don't forget you'd also have to essentially double the human's time if you want them to have the same as before as they can no longer usefully think on the bots minuscule time (of course you have to turn off pondering for this).


Also, I even think the humans chances might get worse, because it's psychologically very hard to think for hours and have your opponent always blitz out an instant reply, which feels like "Ha! I saw that coming".

Re: Predict the winner of World Go Championship 2017

Posted: Thu Mar 09, 2017 1:26 am
by Ulquiorra
gowan wrote:I just wonder how time handicaps could be done. What if computers were really short-timed, such as one second per move? Of course the human players might have five minutes per move. Would that equalize the playing field?


Why are we trying so hard to handicap the computers? If we don't like losing, the humans should just git gud. Besides Zen did not beat Chou Chikun, so it should already be a fair match even if its leveled up a bit.

Anyways, I'll cast my bet for Iyama Yuuta, just for the novelty of having Japan win at Go again.

Re: Predict the winner of World Go Championship 2017

Posted: Thu Mar 09, 2017 5:27 am
by pookpooi
There should be livestream with English commentaries for this event.

Re: Predict the winner of World Go Championship 2017

Posted: Thu Mar 09, 2017 6:10 am
by Baywa
HermanHiddema wrote:Yep, definitely 5 x 1 minute byoyomi.


Here's what I got from the japanese site:

持ち時間3時間、残り5分から秒読み

Google translate:

Mochijikan 3-jikan, nokori 5-bu kara byōyomi
Duration 3 hours, remaining 5 minutes to countdown

For what it's worth...

Re: Predict the winner of World Go Championship 2017

Posted: Thu Mar 09, 2017 10:45 am
by oren
Baywa wrote:Duration 3 hours, remaining 5 minutes to countdown
For what it's worth...


Yes, it's a common Japanese time keeping. For most of us playing online, it's basically 2:55 + 5 x 1m per side.