Tri-color Balancing Go

[yebellz]

There are many “living” or “safe” groups that are not “uncondtionally alive” as you’ve defined it (which I guess is similar to the concept of “pass alive”). For example, a group that has just one eye that consists of 5 points in a line. Or, another example, a group with two eyes: one is a single point and the other eye is a square of 4 points in the corner of the board. Both of these examples are “living” (in typical Go), but are not “unconditionally alive”, since if one were to only pass, the opponent could still capture it. Typically, a group that even surround large areas of territory, but is not divided into multiple eyes, would not be unconditionally alive.

Hence your proposed rule would not prevent the pathologies similar to suicide.

@Samraku

I decided to table this rule as I was writing it, when it occurred to me that a suicide prohibition might actually do what I wanted.

I don’t quite understand what you wanted now. I thought it was established that only banning suicide would not be enough to prevent all sorts of pathological end game, which you wanted to avoid?

[Jhyn]

Sorry if this has been answered before, but I cannot see what prevents the leading player with more territory to keep filling his own territory at the end, while the two other players must pass. Since suicide is prohibited, each player ends up with one point per group, but can also capture groups from the other players and fill the territory to try to become the middle player. This looks like a strange endgame.

[Samraku]

@yebellz

There are many “living” or “safe” groups that are not “uncondtionally alive” as you’ve defined it (which I guess is similar to the concept of “pass alive”). For example, a group that has just one eye that consists of 5 points in a line. Or, another example, a group with two eyes: one is a single point and the other eye is a square of 4 points in the corner of the board. Both of these examples are “living” (in typical Go), but are not “unconditionally alive”, since if one were to only pass, the opponent could still capture it. Typically, a group that even surround large areas of territory, but is not divided into multiple eyes, would not be unconditionally alive.

Hmm, those are some good exceptions I hadn’t thought of. I’ll try and think of a more robust potential rule (if it’s needed) at work today.

@Jhyn

Sorry if this has been answered before, but I cannot see what prevents the leading player with more territory to keep filling his own territory at the end, while the two other players must pass. Since suicide is prohibited, each player ends up with one point per group, but can also capture groups from the other players and fill the territory to try to become the middle player. This looks like a strange endgame.

It’s area scoring, so filling in one’s own territory won’t do anything once dame have been filled. It will be a strange endgame, the question is how strange, and if that strangeness detracts from enjoyment, or adds to it.

[Samraku]

Suicide

Okay, here’s another proposal. It loses some elegance, but I think the trade-off will be worth it if the current ruleset proves to have insufficient power to stymie the silliness.

Suicide is Forbidden.

It is forbidden to make a move such that a group which was pass-alive becomes no longer pass-alive after the move is played.

On any turn, instead of playing a stone of their own color, a player may (but is never forced to) choose instead to place a stone of either opposing color, provided that it would fulfill the following two conditions:

1. It would be a legal move if played by the owner of that color.
2. It causes a group that was previously not pass-alive to become pass-alive.

A looser (in most cases) restriction #2 would be “It forms an orthogonal connection with a stone of like color”, but I’m not sure if there’s much to recommend this restriction over the one I put above.

Balance

The more I think about the pressures in Tricolor Balancing Go and Tricolor Go, the more I think I may have dismissed the basic variant too early. I was probably biased by all the bad press most three player games get, without considering that Go, by it’s nature, may not share those flaws.

A common problem with 3-player chess is that if red and blue start fighting and exchange a few pieces, they are both weakened with respect to green, and thus players are discouraged from engaging the enemy. I think, in Tricolor Go, that if red and blue get into a fight and green stays out of it, the best case scenario for green is that the fight ends evenly and it’s still an even game overall. But if either red or blue win the fight, then yes, green will be in a good position relative to the loser, but may well have taken a severe blow to their winning chances now that another player has had a good result in a fight. Thus, when a fight breaks out between two players, it will make sense for the third player to join in, and fun will be had by all.

One problem which I think Tricolor Go and Tricolor Balancing Go both have in unique ways, is being able to get ahead/becoming completely lost.

In Tricolor Balancing Go, it is in both the low-score and high-score player to maneuver their scores closer together, so that they can get a chance on overtaking the mid-score player. This will result, between strong players, with the score being very close throughout the game, which may make for an unsatisfying end that may feel as if it’s almost up to chance. I think this may be a fatal flaw in the core concept.

In Tricolor Go, on the other hand, the players may be able to develop a meaningful point spread amongst themselves, as everyone will be trying to gain points while reducing everyone else’s. More specifically, the low-score and mid-score players will both be trying to reduce the high-score player’s score, as well as eachother’s score, while the high-score player need only focus on reducing the mid-score player’s score. I expect there would be some pressure for scores to converge, but it would probably be much weaker than in Tricolor Balancing Go. The downside is that the low-score player might reach a stage in the game where it’s just not feasible for them to win, which may result in a king-maker scenario. This may not be as much of a problem considering that the low-score player only has half the attention of the mid-score player (and none of the attention of the high-score player) focused on keeping them down, which may be an incredibly elegant and balanced catch-up mechanism.

All in all, I’d still like to play a few games of Tricolor Balancing Go at the very least, but the simple solution may be best here.

[teapoweredrobot]

@Samraku

All in all, I’d still like to play a few games of Tricolor Balancing Go at the very least,

I’d love to play as it sounds fascinating. I’m not sure I could in practice as it would need to be correspondence pace.
Are there rules about colluding/discussion during the game? I guess the resigning rule means that discussion is permitted between players but I wonder how that might affect things.

[Samraku]

@teapoweredrobot

I’d love to play as it sounds fascinating. I’m not sure I could in practice as it would need to be correspondence pace.

I’d be interested in starting a second game with the rule proposal I listed above to see the difference it makes. In my, Haze with a Z’s, and Gia’s game we are playing correspondence style, so that’s all good.

@teapoweredrobot

Are there rules about colluding/discussion during the game? I guess the resigning rule means that discussion is permitted between players but I wonder how that might affect things.

I have not reached a conclusion on this. The closest I’ve come is considering putting a recommendation in the rules to agree prior to play on the level of discussion permitted in the game. The major options I see are:

1. No discussion or collusion allowed. (Contract Bridge style)
2. Table-talk allowed, but no collusion.
3. All table-talk including but not limited to alliances and discussions regarding the game state are allowed as long as all 3 players are privy to all messages.
4. As with 3, but private conference between 2 of the 3 players is also allowed.

Players could then agree on which of the four they were playing with. I’d personally prefer 1 or 2, maybe 3 depending on the group, but I think there are people who would like each of them.

[meili_yinhua]

My personal inhibition with a 3+ person game is that, with a square-tiled grid, if two players decide that one player is too much of a threat from the beginning of the game, they can always kill that person’s stones. In 3 player it takes a few moves, looking like this in the end:

which is an uncommon problem for most political games… Usually it takes a lot more to take someone out, and in all that time they can try to talk one of the players out of it. This just feels terrible. And the thing is if the player as b is good enough at the game to be considered as a threat, the triangle and square players might feel compelled to do this multiple times. And to further salt that wound, it is likely that no talking is required to get something like this to happen.

Of course, the problem of “who gets the capture” is probably going to be a point of contention between triangles and squares, but even the first 3 triples of moves is enough to put b in a sour mood.

@Samraku

One problem which I think Tricolor Go and Tricolor Balancing Go both have in unique ways, is being able to get ahead/becoming completely lost.

This is a common problem in political games, and it’s called “effective elimination”, and generally the solution there is to either get the eliminated player(s) to resign, or to actually eliminate them, which is why I’m a tad skeptical for the “resignation is only possible if two players agree” ruling. I’d be far more tempted to do the typical “pieces remain on the board but makes no moves” which can lead to some potential seki-like situations where neither player will move to atari because it gives the other player the stones.

It’s an idea that I think could be really nice if it works, but very, very frustrating if not

[yebellz]

I’m not sure if these changes work, but my suggestion is that your proposed variant is trying to do two things:

1. Play Go with 3 players, which I think is already well understood and explored
2. Introduce a “second place” objective, which can create some pathological strategy that bear less resemblance to Go.

I think it should be possible to separate these two. One way to consider your rules for a second place objective work would be to imagine them applied to a two-player game, which I guess you could think of as anti-Go.

[Samraku]

Yeah, I’m growing increasingly doubtful that this idea is salvageable. Still, it was a productive learning experience as to the dynamics of 3-player games and unintended consequences in game design.

I’m still up for a vanilla 3-player game if we get someone else.