Life In 19x19

This is a variant that Nick Bentley and I discussed when talking about his game Blooms. The rules are the same as for Go, with the following differences:

- Passing is not allowed.
- Suicide (of one or more stones) is allowed.
- Your score at any given time is the number of enemy stones that have been removed from the board so far.
- The first player to reach a score of N points wins.
- The ko rule is recommended, but it is not strictly necessary.
- Superko is not used.

I would love to hear your thoughts on this. Has this exact ruleset been suggested before? How do you think it compares to the standard game and to other forms of No-Pass Go?

For large values of N, the game seems to be very similar to regular Go with territory scoring, despite having simpler rules.

luigi wrote:This is a variant that Nick Bentley and I discussed when talking about his game Blooms. The rules are the same as for Go, with the following differences:

- Passing is not allowed.
- Suicide is allowed.

If you don't allow suicide it would be convenient to give one of your stones to your opponent at your turn when you do not have any legal play except self-atari.

- Your score at any given time is the number of enemy stones that have been removed from the board so far.
- The first player to reach a score of N points wins.
- The ko rule is recommended, but it is not strictly necessary.
- Superko is not used.

I would love to hear your thoughts on this. Has this exact ruleset been suggested before? How do you think it compares to the standard game and to other forms of No-Pass Go?

Instead of suicide I have proposed Capture-N where you can give one of your stones to your opponent at your turn. I prefer to have a ko rule when N > 1. OC, superko could be used with large N. Edit: I also think that giving a stone should lift all ko and superko bans, even though the board remains the same. There is no danger of getting into an infinite loop.

I think Capture-N is a fun game. Straight No Pass Go has values that are counterintuitive for a go player and a steep learning curve. I would not play it with suicide. It might be an interesting game for human spectators to watch bots play, however.

For large values of N, the game seems to be very similar to regular Go with territory scoring, despite having simpler rules.

Indeed. As N increases Capture-N approaches regular go with territory scoring and a group tax. IMO, the progression of Capture-1, Capture-2, Capture-4, and Capture-7 would be a fun way to introduce someone to go. The reason for this progression has to do with the number of plays necessary to capture a one-eyed group, which is a triangular number plus 1. The next games would be Capture-11, Capture-16, Capture-22. I am not sure if Capture-N with large N would be as much fun as regular go. But I do like the group tax.

Very interesting thoughts, Bill, thanks for replying.

Bill Spight wrote:Instead of suicide I have proposed Capture-N where you can give one of your stones to your opponent at your turn.

The effect is the same except for the rare cases where suicide of multiple stones is used to win a capturing race, right? If so, then allowing suicide seems more elegant. The placement rule simply becomes "place a stone on any empty point", without restrictions (other than ko). I should perhaps clarify that my rules are meant to allow even single-stone suicide.

I prefer to have a ko rule when N > 1.

Same here. Without it, all ko situations favor the player who is currently leading in captures, which is probably a bad feature.

OC, superko could be used with large N.

I wouldn't recommend that because I see this ruleset primarily as a way to kill cycles without the annoying bookkeeping involved in superko. Since long cycles are rare, having them favor the current leader should be mostly harmless.

Edit: I also think that giving a stone should lift all ko and superko bans, even though the board remains the same. There is no danger of getting into an infinite loop.

I had missed that loose end in my rules, but your suggestion makes perfect sense, as giving a stone (whether directly or by means of single-stone suicide) changes the game state. It's the same as the button in that regard.

I think Capture-N is a fun game. Straight No Pass Go has values that are counterintuitive for a go player and a steep learning curve. I would not play it with suicide.

Isn't Straight No Pass Go simply broken with suicide? Or did you mean to say suicide of more than one stone only?

As N increases Capture-N approaches regular go with territory scoring and a group tax.

Would it be equally correct to say that it approaches stone scoring? If not, what makes it closer to territory scoring with a group tax than to stone scoring?

IMO, the progression of Capture-1, Capture-2, Capture-4, and Capture-7 would be a fun way to introduce someone to go. The reason for this progression has to do with the number of plays necessary to capture a one-eyed group, which is a triangular number plus 1. The next games would be Capture-11, Capture-16, Capture-22.

I like this progression a lot, but is it still meaningful for high values of N? A big enough eye will eventually become two eyes, so there must come a point where further increasing N doesn't enable capturing any further one-eyed groups, right?

luigi wrote:

Bill Spight wrote:I think Capture-N is a fun game. Straight No Pass Go has values that are counterintuitive for a go player and a steep learning curve. I would not play it with suicide.

Isn't Straight No Pass Go simply broken with suicide? Or did you mean to say suicide of more than one stone only?

Yes, I meant suicide of more than one stone.

For instance, group tax aside, a two point Black eye with a White stone inside it is worth 2 points for Black in straight no pass go, while with suicide of more than one stone White can play a stone as a sacrifice and then, if Black does not make it a one point eye, White can keep playing inside it indefinitely. Edit: In straight no pass go the loss from suicide is infinitesimal. White can keep on sacrificing without losing a point. Suicide, even of more than one stone, turns go into a (potentially) infinite game, even with a ko rule.

luigi wrote:

As N increases Capture-N approaches regular go with territory scoring and a group tax.

Would it be equally correct to say that it approaches stone scoring? If not, what makes it closer to territory scoring with a group tax than to stone scoring?

It approaches both, OC.

In Capture-1 territory has an intuitive meaning once the dame filling stage is over as the number of points where you can safely play a stone, after taking the group tax into account. At this point the players can stop play and simply count the score. The same may be true with stone scoring when you add the number of stones already on the board. However, the connection between a play and an empty point of territory is missing with stone scoring.

In straight no pass go you can evaluate territory, which includes dead stones. However, stone scoring is impossible with straight no pass go. For instance, group tax aside, a two point eye is worth 1½ points of territory, but there is no half stone.

luigi wrote:

IMO, the progression of Capture-1, Capture-2, Capture-4, and Capture-7 would be a fun way to introduce someone to go. The reason for this progression has to do with the number of plays necessary to capture a one-eyed group, which is a triangular number plus 1. The next games would be Capture-11, Capture-16, Capture-22.

I like this progression a lot, but is it still meaningful for high values of N? A big enough eye will eventually become two eyes, so there must come a point where further increasing N doesn't enable capturing any further one-eyed groups, right?

Capture-7 is necessary to kill a four point eye. I think beginners need to learn that. They probably need to learn how to kill five and six point eyes, as well, but when it is good to switch to regular go is an open question. I have heard of adding one point to N up to Capture-5, but I like this progression better.

A fully enclosed 7 point eye will typically become seki or ko, at worst. That's why I stopped after Capture-22. When eyes are not fully enclosed the number of stones to kill could be higher, but by time the number of stones to capture has gotten large, why not just switch to regular territory go with a group tax?