Page 1 of 2

CGT Annotation

Posted: Mon Oct 17, 2011 12:29 am
by RobertJasiek
Apart from assigning Left to Black and Right to White by convention, is there any difference between (1|2) and (2|1)? Is one of them, by convention, a number and the other not or is neither a number?

Re: CGT Annotation

Posted: Mon Oct 17, 2011 2:03 am
by perceval
i don't get the question :scratch:
but that fine i probably won t have the answer anyway

Re: CGT Annotation

Posted: Mon Oct 17, 2011 2:13 am
by p2501
perceval wrote:i don't get the question :scratch:
but that fine i probably won t have the answer anyway
CGT stands for Combinatorial Game Theory

http://senseis.xmp.net/?path=CGTPath&pa ... ctionToCGT

Re: CGT Annotation

Posted: Mon Oct 17, 2011 4:20 am
by Sverre
RobertJasiek wrote:Apart from assigning Left to Black and Right to White by convention, is there any difference between (1|2) and (2|1)? Is one of them, by convention, a number and the other not or is neither a number?
p2501 wrote:CGT stands for Combinatorial Game Theory

http://senseis.xmp.net/?path=CGTPath&pa ... ctionToCGT
Robert's question did not use the notation that seems most common on SL. If the curly-bracket notation on that page is what Robert intends, then I'm fairly certain there's a difference between (1|2) and (2|1), since {0|} = 1 and {|0} = -1. Specifically, {1|2} means Black can move to 1 while White can move to 2, while {2|1} means Black can move to 2 while White can move to 1.

If something different is meant, then I am not familiar with the notation he's using.

Re: CGT Annotation

Posted: Mon Oct 17, 2011 7:18 am
by emeraldemon
{1|2} is a number, it's 3/2. {2|1} is not a number, it's "hot", meaning both players want to play there. I've linked to this elsewhere, but there's a nice java program that will calculate CGT values for you, plot thermographs, and do some other cool stuff:

http://cgsuite.sourceforge.net/

{1|2} doesn't make a lot of sense as a go position, it describes a situation where neither player wants to play in a region, because the opponent will gain a point (maybe a kind of very small seki?). Many CGT games force players to make moves in unfavorable regions, so the theory has to handle this (the red-blue hackenbush game from the CGT books, for example). {2|1} is just a standard 1 pt gote.

In a somewhat confusing choice (I think), the "Chilling gets the last point" book uses a technique called "chilling" to turn hot plays into cold ones, I guess because the analysis is more in line with other cool CGT games. Basically you imagine that every move on the board costs you 1 point to make. So {2|1} actually becomes {2-1 | 1+1} = {1|2} ! As I said, I found this very confusing on my first read through that book :(

Hope that helps.

Re: CGT Annotation

Posted: Mon Oct 17, 2011 9:43 am
by Bill Spight
emeraldemon wrote:{1|2} is a number, it's 3/2. {2|1} is not a number, it's "hot", meaning both players want to play there.

{snip}

{1|2} doesn't make a lot of sense as a go position, it describes a situation where neither player wants to play in a region, because the opponent will gain a point (maybe a kind of very small seki?).
Consider this 6x12 board under area scoring. :)
Click Here To Show Diagram Code
[go]$$ Evaluate this board. (Area scoring)
$$ -------------------------
$$ | . . X O O . O O . O X . |
$$ | O X X O . O X O O O X X |
$$ | O O X O O X X X X X O . |
$$ | O . X O X X . X X O O O |
$$ | X X X O X . X . X X O . |
$$ | O O O O X X X X X O O O |
$$ -------------------------[/go]
What about its value under territory scoring? Not Japanese/Korean, but Lasker-Maas/Spight/Ikeda-something-or-other?

If you chill area scoring you get territory scoring. :)

Re: CGT Annotation

Posted: Tue Oct 18, 2011 2:40 am
by flOvermind
RobertJasiek wrote:Is one of them, by convention, a number and the other not or is neither a number?
That has nothing to do with convention, the term "number" in the CGT sense is strictly defined.


A "game" is defined as some {L|R}, where L and R are sets of games.

Let X := {XL|XR} and Y := {YL|YR}. X <= Y if and only if:
There is no xL in XL such that Y <= xL, and
there is no yR in YR such that yR <= X.

A "numeric game" is defined as a game {L|R} where L and R contain only numeric games, and each l in L is strictly less than each r in R.

Two numeric games X and Y are "equivalent" iff X <= Y and Y <= X.

A "number" is defined as an equivalence class of numeric games.


Because 1 < 2, {1|2} (or more accurately it's equivalence class) is a number. And of course 2 < 1 is false, so {2|1} is not a number.
Note that 0, 1, 2 and so on are just abbreviations for certain numbers. 0 := {|}, 1 := {0|} and 2 := {1|}.
If you want you can check the axioms to see that 1 < 2 is really true ;).


See http://en.wikipedia.org/wiki/Surreal_number, this article is a lot better than the articles on senseis. In this article, "games" are called "forms", but it's really the same thing. Conway defined the numbers first, and applied them to game theory later.

Re: CGT Annotation

Posted: Tue Oct 18, 2011 4:09 am
by RobertJasiek
Thank you and emeraldemon!

IMO, Conway's definition of numbers in On Numbers and Games, which I read a couple of years ago, is the most elegant. At university we still learned the traditional ways via, e.g., peano axioms though.

I guess I asked because I thought of {1|2} as being the complement of {2|1} but, of course, Right prefers negative numbers. Do I get it right that -{1|2} = {-2|-1}?

Re: CGT Annotation

Posted: Tue Oct 18, 2011 6:50 am
by flOvermind
RobertJasiek wrote:IMO, Conway's definition of numbers in On Numbers and Games, which I read a couple of years ago, is the most elegant. At university we still learned the traditional ways via, e.g., peano axioms though.
Peano axioms define the Natural numbers. On top of them, you can of course constructively define more types of "numbers", like rational numbers, real numbers, complex numbers and so on.

Conway's "numbers" are entirely different mathematical objects, even though they share the same name and some of their properties (essentially, forming an ordered field).
RobertJasiek wrote:I guess I asked because I thought of {1|2} as being the complement of {2|1} but, of course, Right prefers negative numbers. Do I get it right that -{1|2} = {-2|-1}?
Yes.

Re: CGT Annotation

Posted: Tue Oct 18, 2011 3:31 pm
by Bill Spight
Click Here To Show Diagram Code
[go]$$ Evaluate this board. (Area scoring)
$$ -------------------------
$$ | . . X O O . O O . O X . |
$$ | O X X O . O X O O O X X |
$$ | O O X O O X X X X X O . |
$$ | O . X O X X . X X O O O |
$$ | X X X O X . X . X X O . |
$$ | O O O O X X X X X O O O |
$$ -------------------------[/go]
No bites yet?
Click Here To Show Diagram Code
[go]$$W White first
$$ -------------------------
$$ | 1 . X O O . O O . O X . |
$$ | O X X O . O X O O O X X |
$$ | O O X O O X X X X X O . |
$$ | O . X O X X . X X O O O |
$$ | X X X O X . X . X X O . |
$$ | O O O O X X X X X O O O |
$$ -------------------------[/go]
White wins by 1 pt. (35 to 34).
Click Here To Show Diagram Code
[go]$$B Black first
$$ -------------------------
$$ | 1 2 X O O . O O . O X . |
$$ | O X X O . O X O O O X X |
$$ | O O X O O X X X X X O 3 |
$$ | O 4 X O X X . X X O O O |
$$ | X X X O X . X . X X O 5 |
$$ | O O O O X X X X X O O O |
$$ -------------------------[/go]
Jigo (36 to 36).

If White answers Black's ko threat, White loses.

So the original board has a count of -0.5, and a play gains 0.5 points, which is less than a Japanese dame!

Theoretically you could score the original board as -0.5 by territory scoring, but the only proposed form of go that will give you that score is Button Go. :)

Re: CGT Annotation

Posted: Tue Oct 18, 2011 11:59 pm
by RobertJasiek
How very nice! Very well constructed!

Re: CGT Annotation

Posted: Wed Oct 19, 2011 11:06 am
by Bill Spight
Thanks, Robert. :)

Re: CGT Annotation

Posted: Sat Oct 29, 2011 3:43 pm
by cyclops
flOvermind wrote: Peano axioms define the Natural numbers. On top of them, you can of course constructively define more types of "numbers", like rational numbers, real numbers, complex numbers and so on.

Conway's "numbers" are entirely different mathematical objects, even though they share the same name and some of their properties (essentially, forming an ordered field).
In the number farm what makes Conway's "numbers" less equal than the numbers derived from Peano? And why aren't Conway's "numbers" derivable from Peano? Maybe question1 = question2 !

Re: CGT Annotation

Posted: Sat Oct 29, 2011 6:17 pm
by Bill Spight
cyclops wrote:
flOvermind wrote: Peano axioms define the Natural numbers. On top of them, you can of course constructively define more types of "numbers", like rational numbers, real numbers, complex numbers and so on.

Conway's "numbers" are entirely different mathematical objects, even though they share the same name and some of their properties (essentially, forming an ordered field).
In the number farm what makes Conway's "numbers" less equal than the numbers derived from Peano? And why aren't Conway's "numbers" derivable from Peano? Maybe question1 = question2 !
Conway numbers are special kinds of games. Peano numbers are natural numbers, aka counting numbers. Counting is implicit in Conway numbers, but they are not counting numbers.

Peano numbers are usually augmented to include zero. We can write the Peano numbers this way.

{} = 0
{0} = 1
{1} = 2
***

Conway defines 0 as a game that the second player wins. A player wins if her opponent has the move but has no play. The two players are called Left and Right. The simplest game is one in which neither player has a move. It is 0. We can write it this way.

{|} = 0

The vertical bar separates options to which Left can move (to the left) from options to which Right can move (to the right). Neither player has an option, and hence, no move. Whoever has the move loses.

We also have

{0|} = +1
{|0} = -1

+1 is the game where Black to play can move to 0, while White has no play. -1 is the opposite.

Note that Conway numbers skip the natural numbers as go directly to the integers.

We also have

{+1|} = +2
{|-1} = -2

It is also true that {|+1} = 0 = {-1|} = {-1|+1}. All are second player wins.

For more see Winning Ways, by Berlekamp, Conway, and Guy or On Numbers and Games, by Conway.

----

We can use the Peano numbers to define the integers. One way to do that is like this:

(1,1) = 0
(1,0) = +1
(0,1) = -1

I.e., the integer is the difference between the first Peano number and the second.

Re: CGT Annotation

Posted: Sat Oct 29, 2011 6:44 pm
by cyclops
Thx Bill. I know about Peano and about Conway. But your answer does not explain to me why say complex numbers are completely different mathematical objects from Conways `numbers`. As Flover claims. Both are numbers aren´t they? Wait! Could it be that there is no multiplication of Conway´s numbers?