Life In 19x19

Hi,

I've only recently started playing. I have a question that's been bothering me and distracting me from learning things that are undoubtedly more important.

At the end of the game, with territories totally settled (just imagine 2 lines, 1 white, 1 black, of stones across the board), I have been told by strong players (and believe them) that just randomly putting down a stone into your opponent's territory is pointless because the score comes out the same either way, but I can't understand why.

If in the example above, I put a stone down at 2-2 from a corner in my opponent's territory, isn't he going to have to fill in 4 points of his own territory to kill my stone, and I thus gain 3 points? Is the stone just dead by rule so that my opponent can ignore it? But if he passes (at my club we play where you have to give your opponent a stone when you pass), can't I just keep building up a shape until he eventually has to respond? In just idly playing with stones it seems like my opponent has to spend more stones to capture my dead invaders than I'm going to lose.

I can probably figure out how to put up a diagram if my question isn't clear enough.

Thanks!

Have you already learned about the idea of eyes, and groups that cannot be killed?

If you have a group in your opponent's territory that does not have two eyes, he may decide that it cannot make them, and so is a dead group no matter what you do. He might then just pass.

At the end of the game, both players generally agree if some stones are dead. For your example, you probably believe that your single stone is dead, which is fine. He thinks so too, so he passes. You pass too, because you don't think you can make it live. You then start counting, and take the stone off the board as a prisoner for your opponent.

So, overall, your opponent does not have to spend 4 moves to take the stone off, because unless you can prove it is alive it is only reasonable to agree that it is dead.

Try looking at it this way: stones you play in his territory die, giving your opponent one point, and stones your opponent play fill his own territory, subtracting one point. Assuming you play the same amount of times, it cancels out. For example, if you play the 2-2, he plays the 3-3, then you both stop playing there, it's assumed the 2-2 stone is dead, so the added stones cancel out. In theory, you lose nothing from playing in his territory, as long as he replies to every move. However, if, for example, you play the 1-1 and your opponent passes, and the stone ends up dying anyways, you just gave your opponent one extra point, because of the dead stone. Similarly, if you play the 2-2, then your opponent plays two times to be sure it's dead, you gained a point overall, as your opponent is 'throwing garbage into his backyard', as it were.

As amnal has said while I was typing this, your opponent doesn't have to keep playing moves to solidly capture your stones. Unless you play until your group has two eyes or a ton of eyespace, it's more or less assumed to be dead.

Because of the difference of territory (verses area) scoring, you are correct -- if you do spend the extra moves to capture the stone, then you will lose points.

But as long as each player responds equally, then the resulting score remains unchanged.


The problem is in understanding that its not actually necessary to completely surround and capture these invading stones that won't be able to live on their own anyway. (its a kind of shortcut)

This is why it is often suggested that beginners learn using Chinese rules (area scoring) instead. Then it really doesn't matter how many unanswered moves one spends on "killing" already dead stones -- the final score remains unchanged.

Just think of Japanese rules (territory scoring) as a shortcut to the longer area scoring method. And when using "shortcuts", one must be able to recognize dead stones early enough to not waste any more moves on them. Its all about efficiency.

As the other two replies have said, your opponent isn't obliged to actually capture any stone or group of stones you play into his territory, if it's clear to him that, even with an extra move, you can't make it live (either by making 2 eyes or by capturing some of his stones).

But just to answer this point:

Ortho wrote: But if he passes (at my club we play where you have to give your opponent a stone when you pass), can't I just keep building up a shape until he eventually has to respond?

Each time he passes, you get an extra prisoner, thereby an extra point (assuming you're using AGA rules, which this sounds like), but you're also giving him another prisoner by playing further stones which will eventually be removed as dead - so the final score will be no different from if you had simply ended the game before playing stones which couldn't survive. Under none of the rulesets is it advantageous to play a stone at the end of the game which can't possibly live if the opponent responds correctly.

I assume you've come across the fact that there are basically two different ways of scoring a game of go - "territory scoring" and "area scoring". They look different in practice, but can be shown to either give the same result, or have a difference of only 1 point in net score, in virtually all circumstances. AGA rules were created as a compromise between the two - the final score will always be the same as that under area scoring (and therefore possibly 1 different from that under territory scoring), but allow you to get the same score by counting the way you would under territory scoring. Under area scoring, it's obvious that playing dead stones inside your opponent's territory has no effect - it will remain his territory at the end, and whether an intersection within it is empty, contains one of his stones, or one of your dead stones, makes no difference. Under territory scoring - which is the only one where prisoners count - playing this way can easily cost you points, if your opponent is able to pass several times in response to your plays. The "pass stones" of AGA scoring just mean that he can't gain by passing in this way, as each time he has to give you a further prisoner.

But, if you're a beginner, don't worry too much about the intricacies of the different scoring methods - just play! The advantage of using area scoring (and AGA rules count here) is that you never lose anything (other than having to make quite a lot of tedious routine moves at the end of the game) by actually surrounding and capturing all of your opponent's dead stones. In territory scoring, you do, because you have to play inside your own territory to do this.

EDIT: xed_over has posted while I was typing this, and makes many of the same points. I'll leave my post intact, though, as it makes a few different ones.

yeah, I overlooked the fact that he's using AGA rules -- and thus area rules already (with the convenience of territory counting)

Since you are giving pass stones, then the score remains unchanged.

Edited: I now understand completely how this doesn't make any difference in either scoring method.


Now my mind is free of this. Thanks to everyone!

In territory scoring the rules are technically as follows:

After both players pass, a new phase of the game starts where stones are decided to be dead or alive. If there's a dispute in this phase (you say the group is dead, your opponent disagrees), then you resume play and decide the outcome. If you prove that they are dead, you go back to the board BEFORE the dispute resolution sequence, and count the score. Let's say the board is like this:

Click Here To Show Diagram Code

[go]$$B Before dispute (11-18)
$$ +---------------+
$$ | . . . O O O O |
$$ | . . O . O X O |
$$ | O . O O X X X |
$$ | X O O X . . . |
$$ | X O X X . X . |
$$ | X X . . . W . |
$$ | . . . . . . . |
$$ +---------------+[/go]

White disputes that the marked stone is dead. So you play it out something like this:

Click Here To Show Diagram Code

[go]$$B After dispute (11-18)
$$ +---------------+
$$ | . . . O O O O |
$$ | . . O . O X O |
$$ | O . O O X X X |
$$ | X O O X . . . |
$$ | X O X X . X . |
$$ | X X . . 1 W 5 |
$$ | . . . . . 3 . |
$$ +---------------+[/go]

Now White must admit the stone is dead (you captured it!). Now you go back and count the score from the FIRST diagram, not the second.

Oh, I see! That freezing of the score even takes out the "petulant opponent who won't admit they're dead" problem.

Thanks a lot for that! I had no idea that the scoring and the proving life/death were two different phases.

Ortho wrote:Oh, I see! That freezing of the score even takes out the "petulant opponent who won't admit they're dead" problem.

Thanks a lot for that! I had no idea that the scoring and the proving life/death were two different phases.

Unfortunately, when you're playing on an online go server, then neither does the server (know that there are two different phases)

On the server, there's just one phase. So you either play it out and lose the points, or you don't play it out, and just argue with each other instead.

Oh, by the way... welcome to Go.

xed_over wrote:

Ortho wrote:Oh, I see! That freezing of the score even takes out the "petulant opponent who won't admit they're dead" problem.

Thanks a lot for that! I had no idea that the scoring and the proving life/death were two different phases.

Unfortunately, when you're playing on an online go server, then neither does the server (know that there are two different phases)

On the server, there's just one phase. So you either play it out and lose the points, or you don't play it out, and just argue with each other instead.

Oh, by the way... welcome to Go.

Servers could have two phases, and, IMO, should. (The second phase is called the encore.) One of the easiest to implement is a variant of the Lasker-Maas idea. In the encore award a player one point for each stone played on the board.

Click Here To Show Diagram Code

[go]$$B Encore
$$ +---------------+
$$ | . . . O O O O |
$$ | . . O . O X O |
$$ | O . O O X X X |
$$ | X O O X . . . |
$$ | X O X X . X . |
$$ | X X . . 1 W 5 |
$$ | . . . . . 3 . |
$$ +---------------+[/go]

Black plays three stones to capture the White stone. That reduces her territory by three points, but she is awarded three points, which cancels the reduction out.

Under Lasker-Maas rules the encore is played using captured stones. Clever, eh? BTW, Lasker-Maas rules do not necessarily yield the same results as Japanese or Korean rules.

Bill Spight wrote:Black plays three stones to capture the White stone. That reduces her territory by three points, but she is awarded three points, which cancels the reduction out.

But that just sounds like Pass Stones now

KGS implements AGA rules (but I suspect counts using Area rather than Territory)

xed_over wrote:

Bill Spight wrote:Black plays three stones to capture the White stone. That reduces her territory by three points, but she is awarded three points, which cancels the reduction out.

But that just sounds like Pass Stones now

It is superior to using pass stones because you do not have any forced passes to keep the number of stones even.

Keep it simple. Playing useless stones into your opponents territory is a dick move. Don't be a dick.

Redbeard wrote:Keep it simple. Playing useless stones into your opponents territory is a dick move. Don't be a dick.

Unless, of course, you're not sure if the stones are useless, or there might be a way to live. If you honestly can't tell, then I definitely recommend you go for it. Often, your opponent will have a lot of cutting points and can be successfully invaded 100% of the time, even if you can't read out the sure way to live/kill before you start. I'm stating what may not need it, but I'm a master of the obvious. =)