Axd/Rules

Sub-page of Axd

IMO, the driving force behind Go is the beauty of the game (eg the Japanese notion of emptyness[1], then its relation between wood, lines, black and white): it should be reflected in the rules.

I always try to convince novices around me about the simplicity of the game; but in reality, the surface below is darker.

Table of contents

Dark areas

I usually introduce the game using stone counting. Sometimes I will switch to chinese rules, to end with japanese counting.

Sunjang Baduk (without the initial setup) looks like a logical next extension to Go due to this new phase added: removing stones. Seen the complex rulesets that appear (most of this is hidden from the casual player), maybe we should go back to the roots.

Root concept: owned area

Most rules define area as "completely enclosed" by own color. Enter Proximity Scoring...

Dead groups

[thumbnail diagram]

One of the problems seems to come from the counting phase, where groups are to be accepted as dead (or alive) - but it might not always be clear. See also Alternating Play.

Need to review this paragraph, it is getting totally blurred...


[Diagram]

counting phase (beginners): 6 points for Black (Chin.)

Another issue is that strictly speaking, novice players can leave groups in atari on the goban and call them alive; I don't think there is a fundamental problem here, but it is unusual, and more experienced players could intervene in such a situation - but is it necessary?


Minue is not the first to state that Go is not a game of territory ("It can make a beginner have a wrong idea that Go is a game of territory"); yet I can't grab the idea behind it. The result IS about counting; maybe Go is about creating strong groups so that one can acquire territory/area, but the end result IS something two-dimensional, IMO. Maybe I confuse goal with means to achieve it. Maybe this confusion should be cleared out elsewhere.

This is another unclarity in the rules: most games have a goal, in Go is it to be the one with the most points, measured in some form of surface won. But different rules exist that say how to count; the outcomes (usually) are (slightly) different, that's the problem.

Some links on this subject:

BenjaminTeuber / Guide To Become Strong
where the introduction of Japanese rules could have created this split
And then there are beasts like Disturbing Kos, Superko Anomalies...
See also HermanHiddema/Local Global Relations In Rules Disputes
MoonshineLife
1-Eye-Flaw
Anti Seki
HermanHiddema/Go and Chess, a comparison
Thus go requires that a player understands some very basic tactics before really understanding the goal of the game. A sort of catch 22 of the rules of go.

Handicaps

Handicaps versus grade difference is another point of question: first, I don't see why getting an amount of handicap stones can guarantee an equitable game. For Dan players, this is no longer true; so I suspect that this is only a very rough rule-of-thumb to more or less compensate the weaker player. But I see no alternative; it does seem to work, though the mechanism is not quite clear to me.

And why the fixed setup (hoshi, handicap placement)? It just adds almost never used stuff to the rules. And it does not seem to help the beginner, as the more experienced player knows his/her way around the handicap stones. OTOH, having a fixed setup provides a well-known framework for the weaker player.

Recently I extended the tables at handicap for smaller board sizes because until a snappy formula can pack all this, it it nice to have a printout of these tables when playing against far weaker players. But why would there be a linear relationship between H and rating diff?

And the fact that there are different ways to assign H/K values is another indication that this system is somewhat buggy.

I think that a good way to determine these values is to tap this huge database that is the collection of go servers: they all have loads of played games in various sizes and settings, it must very well be possible to

  • confirm the current komi (6.5, or any other...)
  • determine probable H/K values that guarantee even games. this must be statistically feasible, just like the way the current komi (whatever it is ;-) was determined.

So following should be taken into account: B rating, W rating, H, K, outcome, size, #moves.

  • outcome to filter away resigns, loss by time, ...
  • #moves to filter away silly games

Note that this is only slowly maturing in me; I'm aware that other people may already have walked this path and have found the answers I'm still looking for.

Suicide

I think that suicide should be permitted. It gives extra possibilies, and imposing a no-suicide rule adds more ballast (and HardHeartedness) to the rules than adding to the beauty of the game. Still need to figure out how to deal with the match clock when playing a single-stone suicide (multi-stone suicide is followed by removal and then pushing the clock).

To develop: I tend to see no difference between a suicide move and a pass stone; as a result, this makes the case of suicide moves more plausible. (vague link with No Pass Go/Suicide)

Restart from scratch

So how about starting from the most primitive rules possible, and transform the rules step by step, on the condition that the outcome is guaranteed to remain the same in every step? Clock related issues are not central to rules: instead, such rules should be made "modular" (read: independent of the rules chosen).

Currently New Zealand Rules and AGA Rules sounds interesting.

Links


[1] the "beauty of omission", (a criterion possibly more appropriate to Noh opera than to a worldwide game of strategy), [ext] Bill Taylor, John's Go page


This is a copy of the living page "Axd/Rules" at Sensei's Library.
(OC) 2009 the Authors, published under the OpenContent License V1.0.
[Welcome to Sensei's Library!]
StartingPoints
ReferenceSection
About