Yesterday I was experimenting with some cycles when I noticed that all the ones that I know of (basically the ones mentioned in the Sensei's Library entry) could be easily prevented by replacing any ko rules with the following, which I'm calling the "passive ko rule":

If your opponent has just made a capture on his turn, you can't make a capture on your turn.

Of course, the game would be different with this rule, but not by much. Interestingly enough, ko fights proceed almost the same as with normal rules, with the difference that you can't make a capture as a ko threat. On the other hand, snapbacks are not possible.

Is there any position where optimal play would lead to a cycle under this rule? I can't find any, but maybe they are possible. Anyway, the ratio of void games must be necessarily lower than it is under basic ko rules by many orders of magnitude...

Also, do you know if this rule has been mentioned elsewhere?

EDIT:

The following rule preserves snapbacks while still preventing all known cycles:

If your opponent has captured n stones on his previous turn, you can't make a capture on your turn unless it's bigger than n stones.

Related ko rules:

http://home.snafu.de/jasiek/koclass.html

Your variant is a new idea, AFAIK.

It certainly fixes cycle problems, but making snapbacks illegal would make it a very different game to play.

That sounds pretty similar to the rules I remember reading somewhere on go in Tibet, where aside from playing on a 17x17 board, the ko rule would apply to any capture. However it would only apply to the capturing stone, not the whole board.

The Tibetan ko rule is: you cannot play a stone on a place where your opponent just took one of your stones

Perhaps even bigger than snapbacks is that a great many trades become impossible.

_________________
Occupy Babel!

Click Here To Show Diagram Code

[go]$$Bc
$$ ------------------
$$ | . . . . . . . . .
$$ | . . . . O O . . .
$$ | . O O O O X . X .
$$ | O O O X O X . . .
$$ | X X X a X X . . .
$$ | . . X . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

you couldn't play a as a ko threat.

And some beautiful under-the-stones plays are illegal too.

_________________
Help make L19 more organized. Make an index: https://lifein19x19.com/viewtopic.php?f=14&t=5207

That's kind of crazy. It means

Click Here To Show Diagram Code

[go]$$Bc
$$ ------------------
$$ | . O O . X O . . .
$$ | X X X X X O . . .
$$ | O O O O O O . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

is actually seki. This rule completely changes life and death.

Under tibetan rules, is the following alive, dead, or in seki (assume it's black to play)?

Click Here To Show Diagram Code

[go]$$Bc
$$ ------------------
$$ | . . . . . . . . .
$$ | . O O O O O O . .
$$ | O O X X X X O . .
$$ | O X X . . X O . .
$$ | O X . O . X O . .
$$ | O X X X X X O . .
$$ | O O O O O O O . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

This also seems like Seki to me.

You're right, sorry. it is only illigal if you are capturing in a place where your opponent took one of your stones.

Just for the sake of comparison, my suggested variant has much more to do with standard Go than with Tibetan Go. Lamp, your examples are indeed seki under Tibetan rules, but they behave exactly as in Go with the passive ko rule.


Actually you were right the first time.

That's indeed an important difference, but I think such trades are actually much less frequent than one might expect.

In fact, I've reviewed a fair bunch of pro games from my small database of 5000 games (included with the SmartGo demo version) in search of moves that would have been illegal with the passive ko rule (from now on divergent moves) and I've observed that not only are there just about two or three such moves per game on average, but also they're usually part of ko fights, i.e. captures made as ko threats or ko captures made after a capturing reply to a ko threat. My rough estimation is that 80% of divergent moves are part of ko fights, and the remaining 20% are trades of the kind you mention. Also, about 30% of the moves which are part of ko fights are divergent moves.

Out of curiosity (or maybe just to play devil's advocate against myself), I've also reviewed the longest game from my database, a 364 moves struggle from 2006 between Chang Hao 9p and Wang Lei 5p, and found 10 divergent moves. Surprisingly enough, all but one of them were part of ko fights, and even the remaining one was a capture of a single stone immediately after a ko capture by the opponent. It's also interesting to note that all divergent moves took place well into the endgame: the first was the 219th move, and half of them took place in the last 26 moves of the game.

All in all, my modest research seems to indicate that using the passive ko rule instead of the basic ko rule would have three main practical consequences:

a) Ko threats would be about 30% less abundant, and ko fights would be shortened accordingly.

b) There would be a slightly higher incentive to make captures.

c) Cycles would be either impossible or astronomically rare.

No, neither of these positions are seki.
To clarify: The Tibetan Ko rule states that you cannot place a stone where a capture stone has just been removed from.

Lets say we have A then B (capturing). This is now a ko, not a seki.

If you actually play a game under Tibetan rules, you might get a better feeling for this. I have done, I must say that the gameplay wasn't significantly different.

_________________
North Lecale

Well what i meant is that if white realizes he cannot win the ko, he will just not play A to begin with and then it becomes seki since it's white, not black, that controls whether the ko will be fought

Well, not exactly a ko, since it's necessarily settled after White's ko threat.


OK, all the better then. Passive ko is much less restrictive than Tibetan ko and doesn't change the outcome of any basic nakade shapes, so if Tibetan Go isn't significantly different from Go, playing with passive ko instead must be practically indistinguishable from Go.

A lot of the 'I capture 2 (or 3 or 4), you capture 1' positions are totally different with this rule.

Passive ko sounds interesting. I think that both it, and Tibetan ko, could be used with beginners.

Click Here To Show Diagram Code

[go]$$Wc
$$ ------------------
$$ | 1 O O 2 X O . . .
$$ | X X X X X O . . .
$$ | O O O O O O . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

So this is how I understand the situation. White makes the internal atari play, Black takes.

Click Here To Show Diagram Code

[go]$$Wc
$$ ------------------
$$ | S S S 2 X O . . .
$$ | X X X X X O . . .
$$ | O O O O O O . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . 3 . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .[/go]

White now cannot break the eyespace, so plays a ladder breaker as a ko threat. Black can answer the threat, or he can live.

_________________
North Lecale

1) If there are two to three moves ruled out by this, then there will be several times as many places where the best play is different, because illegal moves were present as threats.

2) Two to three moves that can't be played is already a lot!

_________________
Occupy Babel!

??? how do you know this? it seems like apples and oranges to me, but even so I can't help but feeling that yours is more restrictive.

Tibetan ko turns lots of standard l&d into ko (bulky eye spaces and so on). Luigi ko hugely changes situation where multiple strings of stones are in Atari, and turns a lot of them into ko. Both very broad, but different.

Cycles are a fly on the window. The ko rule is a fly swatted. Anything more drastic is a brick.