Home » Forum Home » General

Topic: Go
Replies: 7   Views: 65,158   Pages: 1   Last Post: Jan 18, 2018, 9:17 PM by: watsu

Search Forum

Back to Topic List Topics: [ Previous | Next ]
Replies: 7   Views: 65,158   Pages: 1  
rainwolf

Posts: 763
Registered: Apr 12, 2008
From: Singapore
Age: 44
Home page
Go
Posted: Jan 16, 2018, 9:53 PM

I'm working on Go for Pente.org, and as I'm not familiar with the game, I have a few questions.

To avoid overly cross-referencing posts in this thread, I'm going to just do one question at a time.

First one: is there an upper bound on the total number of captures in a game per player?


rainwolf

Posts: 763
Registered: Apr 12, 2008
From: Singapore
Age: 44
Home page
Re: Go
Posted: Jan 18, 2018, 1:40 PM

For now I'm assuming games will be at most 361 moves per player.

Next question: How do you start a game? Or more specifically, how do you start a game with handicaps?

Since nothing is bootstrapped yet when it comes to Go-ratings, I'm assuming all games will be started without a handicap and with a 7.5 komi, right?

Are handicap stones then only reserved for unrated games? Or do they also make sense in rated games?

watsu

Posts: 1,461
Registered: Dec 16, 2001
Home page
Re: Go
Posted: Jan 18, 2018, 3:45 PM

Note: I like to research and am interested in Go, rating systems and so on, but am NOT (so far) a Go player, so take my answers below with a grain of salt.

Handicap stones placement reference link:
https://senseis.xmp.net/?OrderOfPlacementOfHandicapStones

To attempt to answer the question of handicap stones and ratings, I believe that it would be theoretically possible to incorporate handicap stones into a rating system, but likely needlessly complicated for practical purposes. If ratings were combined with handicap stones, ratings would remain nearly unchanged in the event of a close to expected outcome given the handicap - a game won or lost by a few points, but would change if one player performed significantly better or worse than the expected result given the handicap - more of a blowout than expected.
I would not (at least initially) attempt to implement ratings in combination with handicap stones, in other words. They can be used for unrated games between widely different rated players to make a game between them more interesting, but calculating how, how much and when to change ratings for handicapped games would be a tough call.


Message was edited by: watsu at Jan 18, 2018 4:01 PM


Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
watsu

Posts: 1,461
Registered: Dec 16, 2001
Home page
Re: Go
Posted: Jan 18, 2018, 3:55 PM

First one: is there an upper bound on the total number of captures in a game per player?

So far, I don't have a great answer for you on this question. My best stab at it is that since a few closely contested professional games will last for 400 moves, but most games will last under 400 (or even 200) moves one could likely assume a tentative upper bound of 200, but I would try to incorporate some sort of capture overflow system beyond any reasonably set upper bound, since players here have been known to accidentally crash the system while recreating with possible but unrealistic game scenarios, specifically with continuous simultaneous poof captures.

Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
rainwolf

Posts: 763
Registered: Apr 12, 2008
From: Singapore
Age: 44
Home page
Re: Go
Posted: Jan 18, 2018, 5:38 PM

Handicaps are for later then, should be a mild adjustment.

The game is over when 2 players pass, but then what?
My guess is the flow is as follows:
The 1st player to pass gets to mark dead stones, the 2nd player can agree or disagree.

With dead stones marked and territory corrected, a final score is computed and the game decided. But what if the last player to pass does not agree? Game continues?

watsu

Posts: 1,461
Registered: Dec 16, 2001
Home page
Re: Go
Posted: Jan 18, 2018, 6:13 PM

I found this link to be useful for score counting - it is for 9x9 games but presumably would hold true for 19x19 as well.

http://cugosoc.soc.srcf.net/go/rules_06.html

Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
rainwolf

Posts: 763
Registered: Apr 12, 2008
From: Singapore
Age: 44
Home page
Re: Go
Posted: Jan 18, 2018, 8:23 PM

I think I know how to count, but I'm wondering about the flow.

I can do the counting by doing a flood-fill for each color and discard the areas that can be flooded by both. But, areas without dead stones removed will be discarded, and with dead stones removed would be counted towards someone's territory.

I don't think there's an algorithm to remove the dead stones (other than playing out the game), I thought players have to agree on them.

So my question is, what if they don't agree? Does play resume to settle the dispute?

watsu

Posts: 1,461
Registered: Dec 16, 2001
Home page
Re: Go
Posted: Jan 18, 2018, 9:17 PM

Okay, I read up a bit more on your question and it seems you are correct that play resumes if there is disagreement on whether or not a stone or stones are dead. There is some discussion of that here: https://boardgamegeek.com/thread/84170/removing-dead-stones

Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
Replies: 7   Views: 65,158   Pages: 1  
Back to Topic List
Topics: [ Previous | Next ]


Powered by Jive Software