Ultimate Go - very simple, yet solving the free teire issue

[Yoav Yaffe]

I would love to get feedback on a proposed ruleset called "Ultimate Go" (see [ https://senseis.xmp.net/?UltimateGo ]) which is:

(1) Extremely simple - actually based (as I found out at some point) on "no-pass Go with prisoner return", so e.g. "score" is only implicit from the rules.

(2) Solves the "free teire" issue, i.e. at the end of a tight game players don't get to play reinforcement moves "for free" (regardless of dame parity). By that I mean every such teire move *will* affect the "score", or more precisely will affect the win\lose status of the game given any updated "Komi".

The solution via the device of an "ultimate prisoner", which signals the transition to an unofficial "counting phase" of the game, introduces only a slight complication to the rules.

[Elom0]

Couldn't you solve the group tax issue by changing 'declare a draw' to 'declare a tally', which counts the number of places any opposing side can't play on their next move? The one who has the higher tally wins. With this it seems like Ultimate baduk rules really are the best around! Each players supply can be anywhere that's not the prisoner area which could consist of the lids and the clock, making the practical application of the ruleset more elegant than AGA rules (place the ultimate prisoner on the clock, and the other delays anywhere outside the clock or lids). Instead of the superko rule, after 360 moves, the definition of supply switches to only the prisoner bowl.

[Joaz Banbeck]

Despite having played go for over 30 years, I did not know the term 'teire'. For those of similar ignorance, from SL:

Te-ire (手入れ) is a Japanese Go term referring to a play which reinforces, repairs, or fixes up one's shape.

[Yoav Yaffe]

Couldn't you solve the group tax issue by changing 'declare a draw' to 'declare a tally', which counts the number of places any opposing side can't play on their next move? The one who has the higher tally wins. With this it seems like Ultimate baduk rules really are the best around! Each players supply can be anywhere that's not the prisoner area which could consist of the lids and the clock, making the practical application of the ruleset more elegant than AGA rules (place the ultimate prisoner on the clock, and the other delays anywhere outside the clock or lids). Instead of the superko rule, after 360 moves, the definition of supply switches to only the prisoner bowl.

I don't believe the "group tax" (which is *implicit* from the Ultimate Go rules) presents a real issue, hence I don't wish to solve it... I think that group tax makes a lot of sense (a group with two one-intersection eyes doesn't really have territory, does it?), anyhow if a simple ruleset implies something non-traditional, I prefer it to a complicated ruleset which justifies some tradition. The same logic which dictates "no territory in a Seki" also dictates group tax IMHO.

[CDavis7M]

I don't believe the "group tax" (which is *implicit* from the Ultimate Go rules) presents a real issue, hence I don't wish to solve it... I think that group tax makes a lot of sense (a group with two one-intersection eyes doesn't really have territory, does it?), anyhow if a simple ruleset implies something non-traditional, I prefer it to a complicated ruleset which justifies some tradition. The same logic which dictates "no territory in a Seki" also dictates group tax IMHO.

I agree that so-called "group tax" is not an actual issue and that the rules of Go are better without additional rules to prevent group tax.

Even the "simple" Ultimate Go rules contain plenty of extraneous rules. It's better game design for rules to leave much of this to the players to decide themselves. Some of the other rules just add complexity for little or no gain. If the idea is to create "simple" rules then there's no need to:

• require black and white
• require identical stones
• have a prisoner bowl
• have an endless supply of stones
• require 19x19
• require that the board be empty at the start
• but a stone in the prisoner bowl at the start
• Black go first
• require a "placement" to be "legal."
• have a rule on "delay" (prisoner giving)
• Have prisoners at all
• Declare a draw

The only necessary rules are:
(1) Play: the players take turns (a) placing an object on an intersection of a grid and (b) removing objects.
(2) Remove: a group of objects connected along grid-lines is removed from the board if not connected to an empty intersection. Optional Ko rule.
(3) Victory: the player having the most stones on the grid wins.

Still simpler, it is the removing of game-pieces that is the only novel aspect of Go and the only rule that is "needed." Alternative play on intersections and scoring are inherent to the removing rule and general game design and can be left up to the players.

The players together can agree about any komi or group tax themselves. The players can decide what object to use, which color or colors they want to use and whether that it a problem for them or their opponent. Tracking and giving prisoners does not make much of a difference. So is banning suicide. I don't even think there is an issue with removing the play that fills the opponents one eye from the board. Even the Ko rule could be optional or "advanced" rule. The players are also free to agree to these things. But requiring them in the rules moves the rules away from "simple."

[RobertJasiek]

Even the Ko rule could be optional or "advanced" rule.

For what purpose? Allowing infinite repetition does not create a playable game.

[Elom0]

Even the Ko rule could be optional or "advanced" rule.

For what purpose? Allowing infinite repetition does not create a playable game.

Because if you're going to create any ko rule, the question comes up of what nature of any superko rule would be like, situational or not, and it's more elegant to include no ko rule and leave everything ko to the players, rather than include a ko rule but just leave super ko to the players.

It's not obligatory for the ruleset to prevent infinite games. What if God wanted to play? She'd have no problem playing go forever.


My amendment isn't any less simple, but naturally doesn't involve group tax.
The only necessary rules are:
(1) Play: the players take turns (a) placing an object on an intersection of a grid and (b) removing objects.
(2) Remove: a group of objects connected along grid-lines is removed from the board if not connected to an empty intersection. Optional Ko rule.
(3) Victory: the player having the most stones intersections the opponent(s) can self-capture with on the grid wins.
And once you have over 350 stones on the board, it's much easier to count eyes.


However technically they are more likely:
(1) Play: the players take turns (a) placing an object on an intersection of a grid and (b) removing objects.
(2) Remove: a group of objects placed by the same player connected along grid-lines is removed from the board if not connected to an empty intersection connected only to objects of a different single player. Optional Ko rule.
(3) Victory: the player having the most stones intersections the opponent(s) can self-capture with on the grid wins.Optional addition and reduction of score based on removed or preremoved objects
And once you have over 350 stones on the board, it's much easier to count eyes.

[RobertJasiek]

Because if you're going to create any ko rule, the question comes up of what nature of any superko rule would be like, situational or not, and it's more elegant to include no ko rule and leave everything ko to the players, rather than include a ko rule but just leave super ko to the players

Elegance is the wrong word. It can mean different things. You might mean minimality.

Just because there are different options for various rules (Black or White starts, with or without suicide, different ko / restriction rules etc.) one must not shy away from choosing some option at least as a default.

A purpose of rules is to allow play under known conditions between different players. Without ko rule, almost every game would lead to a dispute what to do on repetition. Therefore, it is essential to restrict repetition entirely or at least almost always in practice.

[Elom0]

Because if you're going to create any ko rule, the question comes up of what nature of any superko rule would be like, situational or not, and it's more elegant to include no ko rule and leave everything ko to the players, rather than include a ko rule but just leave super ko to the players

Elegance is the wrong word. It can mean different things. You might mean minimality.

Just because there are different options for various rules (Black or White starts, with or without suicide, different ko / restriction rules etc.) one must not shy away from choosing some option at least as a default.

A purpose of rules is to allow play under known conditions between different players. Without ko rule, almost every game would lead to a dispute what to do on repetition. Therefore, it is essential to restrict repetition entirely or at least almost always in practice.

I agree writing 'Optional Ko' is bad. In addition to your reason, Ko is a word that simply means 'forever' and therefore must be defined in the ruleset. I think CDavis7M meant something more like:

(1) Play: the players take turns (a) placing an object on an intersection of a grid and (b) removing objects.
(2) Remove: a group of objects placed by the same player connected along grid-lines is removed from the board if not connected to an empty intersection connected only to objects of a different single player (this is better better for multiplayer games, and a simple ruleset shouldn't assume two players). Optional Ko. Players agree beforehand on how to handle repeated board positions.
(3) Victory: the player having the most stones intersections the opponent(s) can self-capture with on the grid wins. Optional addition and reduction of score based on removed or preremoved objects (this gives the option of territory scoring).

And my standard agreement on how to handle repetition would be:
1) you can't repeat the previous board position 2) each player can place as times as the number of intersections on the board divided the number of players, then plus the number of their objects they have previously placed that have been removed (in practical terms this is better than any superko rule).




If Group Tax should be considered natural if no territory in mutual life is considered natural, then free tiere should probably also be considered natural.

Although I now see free tiere as natural, group tax somewhat natural, but no territory for a group if it happens to be involved in mutal life unnatural.

[Elom0]

(1) Play: the players take turns (a) placing an object on an intersection of a grid and (b) removing objects.
(2) Remove: a group of objects placed by the same player connected along grid-lines is removed from the board if connected only to objects of a different single player. Players agree beforehand on how to handle repeated board positions.
(3) Victory: the player whose objects connect to the most intersections the opponent(s) can self-capture with wins. Optional reduction of score based on removed or preremoved objects.

The standard resolution of repetition is:
1) you can't repeat the previous board position 2) each player can place an object as many times as the number of their objects they have previously placed that have been removed, plus the number of intersections on the board divided the number of players

[CDavis7M]

Even the Ko rule could be optional or "advanced" rule.

For what purpose? Allowing infinite repetition does not create a playable game.

There are many games that have the option of a stalemate which are not just playable but which are great games.

The purpose of an optional Ko rule would be to make the rules simpler. A player might only play a particular board game 3 or 4 times, or even just once. Some people never even finish their first game because of complicated rules. Go can be played a hundred times over tens of hours without the a ko making a difference (one player would decide to play elsewhere). Surely the Ko rule is a great rule for advanced players but it's an advanced rule and not essential to the game.

[CDavis7M]

(3) Victory: the player having the most stones intersections the opponent(s) can self-capture with on the grid wins.

Personally I think it's simplest if the scoring mechanism is directly tied to the play mechanic - play a stone get a point (Have your stones removed, lose points). Relying on another rule like "self-capture" is more complicated.

[CDavis7M]

I agree writing 'Optional Ko' is bad. In addition to your reason, Ko is a word that simply mea...
Optional Ko. Players agree beforehand on how to handle repeated board positions.

No, I think Go is still a good game without a Ko rule, especially for beginners in their first 3-4 games, even in their first 100 games. There is no need for the players to decide about something so rare and unlikely in a beginner game. Advanced players can agree beforehand to an optional ko rule. But beginners would just figure it out during the game, either to drop the game, or one of the players might decide to play elsewhere, or the players might remember there is an optional advanced rule that they are ready for.

Anyway, my point is that the "Ultimate Go" rules claim to be simple but they are actually more complicated in some ways and the goal of handle the "issue" of tree teire is really a non-issue. Teire should be freely playable in a game. Free-teire being less artistic is a different topic.

If the goal is to simplify Go then only 1 rule is needed, the "surrounding" rule. The play mechanic is described in the surrounding rule and alternate play and scoring based on the play-mechanic are inherent parts of board gaming. The Ko rule is an excellent rule, but it's a rule for advanced players. I'm just talking about the normal ko rule, not super-ko or any of its alternatives. Those are not good game rules, they are only helpful/fun for mathematical completionists.

[RobertJasiek]

The purpose of an optional Ko rule would be to make the rules simpler. A player might only play a particular board game 3 or 4 times, or even just once. Some people never even finish their first game because of complicated rules. Go can be played a hundred times over tens of hours without the a ko making a difference (one player would decide to play elsewhere). Surely the Ko rule is a great rule for advanced players but it's an advanced rule and not essential to the game.

I see the point you are making. However, for such occasional players, a rule "repetition is prohibited" is not too much to read. The rules would still be much shorter than for almost all other games.

Relying on another rule like "self-capture" is more complicated.

If it is formulated as an extra concept, it becomes more complicated regardless of whether the rule is with or without suicide. The move rule does need either condition, so for the sake of defining a play, with or without suicide are equally simple / complicated.

so rare

Sorry, but your advertisement of rarity is an exaggeration.

Advanced players can agree beforehand to an optional ko rule.

It can be simpler than that: put "advanced" rules in an initially hidden section. There is no need for having to agree on what the rule is - it is sufficient to look it up as soon as necessary. E.g., the specification of the starting player's colour can be put to the advanced rules.

If the goal is to simplify Go then only 1 rule is needed, the "surrounding" rule.

Haha. No. There must also be the alternation rule to start with. A game aim also won't hurt. After all, go is not just collaborative art painting.

alternate play and scoring based on the play-mechanic are inherent parts of board gaming.

LOL. You make too many assumptions. Games are not per se alternating and not per se scored at all, let alone scored in a particular manner. Go is about points - not about making the nicer contribution to an art painting. Therefore, alternation and score have to be specified by the rules.

I'm just talking about the normal ko rule, not super-ko or any of its alternatives. Those are not good game rules, they are only helpful/fun for mathematical completionists.

Complete description of a game is good game ruling for everybody - not just for logicians.

[Elom0]

I agree writing 'Optional Ko' is bad. In addition to your reason, Ko is a word that simply mea...
Optional Ko. Players agree beforehand on how to handle repeated board positions.

No, I think Go is still a good game without a Ko rule, especially for beginners in their first 3-4 games, even in their first 100 games. There is no need for the players to decide about something so rare and unlikely in a beginner game. Advanced players can agree beforehand to an optional ko rule. But beginners would just figure it out during the game, either to drop the game, or one of the players might decide to play elsewhere, or the players might remember there is an optional advanced rule that they are ready for.

Anyway, my point is that the "Ultimate Go" rules claim to be simple but they are actually more complicated in some ways and the goal of handle the "issue" of tree teire is really a non-issue. Teire should be freely playable in a game. Free-teire being less artistic is a different topic.

If the goal is to simplify Go then only 1 rule is needed, the "surrounding" rule. The play mechanic is described in the surrounding rule and alternate play and scoring based on the play-mechanic are inherent parts of board gaming. The Ko rule is an excellent rule, but it's a rule for advanced players. I'm just talking about the normal ko rule, not super-ko or any of its alternatives. Those are not good game rules, they are only helpful/fun for mathematical completionists.

Sorry! I meant to write Players can agree beforehand, but I was being sloppy . . .

Actually you could replace the ko rule with total move limit. Not the Go we're used to but still Go.

1-the players take turns placing an object on an intersection.
2-A group of objects placed by the same player connected along grid-lines is removed from the board if connected only to objects of a different single player. Repeating the most recent position may be disallowed.
3-If it wouldn't be possible for a player to place two objects without removing one their objects, the game ends. The winner is the player whose objects connect to the most intersections connected only to objects they placed minus the number of their objects that were removed. Each player can place an object as many times as the number of their objects they have previously placed that have been removed, plus the number of intersections on the board divided the number of players.