Ko research and theory

RobertJasiek · Post by **RobertJasiek** » Mon Jan 24, 2011 9:44 am

Some endgame kos are solved:

http://home.snafu.de/jasiek/kodame.pdf

Magicwand · Post by **Magicwand** » Mon Jan 24, 2011 1:12 pm

RobertJasiek wrote:Some endgame kos are solved:

http://home.snafu.de/jasiek/kodame.pdf

i guess you took long time creating that paper but as usual it is whole lot of word of nothing.

what benefits do you think people will have by reading your pdf?
do you think it will improve their endgame?
what level of players should read this?
how about summary on what to expect on your pdf?

RobertJasiek · Post by **RobertJasiek** » Mon Jan 24, 2011 1:47 pm

Magicwand wrote:i guess you took long time creating that paper

Ca. 3 weeks for the mathematical proofs and 2.5 weeks for writing + correcting the paper.

but as usual it is whole lot of word of nothing.

Where nothing is worth up to 4 points of the final score. That's more than the 1 point of Mathematical Go Endgames;)

what benefits do you think people will have by reading your pdf?

Obvious.

do you think it will improve their endgame?

If they did not know how to fight dame ko fights, sure.

what level of players should read this?

All levels of players, who play at least some games in their life under area or stone scoring. All interested in mathematical go theory anyway.

how about summary on what to expect on your pdf?

The PDF itself has good summaries.

Bill Spight · Post by **Bill Spight** » Mon Jan 24, 2011 2:57 pm

Magicwand wrote:
RobertJasiek wrote:Some endgame kos are solved:

http://home.snafu.de/jasiek/kodame.pdf

i guess you took long time creating that paper but as usual it is whole lot of word of nothing.

what benefits do you think people will have by reading your pdf?
do you think it will improve their endgame?
what level of players should read this?
how about summary on what to expect on your pdf?

Robert explores some not uncommon so-called "half point ko" situations, some of them where one player is komonster, who can not only win the ko, but do so when the ko is the last play. Such situations have caught pros unawares who do not usually play by area scoring.

If one player is komonster at territory scoring, the "half point ko" can act like this kind of play:

Click Here To Show Diagram Code: [go]$$ $$ - - - - - - $$ . X . . O . $$ . X X X O . $$ . . . . . .[/go]

Under area scoring it can act like this kind of play:

Click Here To Show Diagram Code: [go]$$ $$ - - - - - - $$ . X O . O . $$ . X X X O . $$ . . . . . .[/go]

Robert also explores cases where there are multiple "half point kos".

Time · Post by **Time** » Thu Jan 27, 2011 11:52 am

Magicwand wrote:
RobertJasiek wrote:Some endgame kos are solved:

http://home.snafu.de/jasiek/kodame.pdf

how about summary on what to expect on your pdf?

Expect to enjoy reading it if you care about some formal go theory.

Expect to put in more effort than you ever imagined to gain 1 point in 5% of your games otherwise. It's obviously not an efficient way to improve at go.

RobertJasiek · Post by **RobertJasiek** » Thu Jan 27, 2011 12:22 pm

Time wrote:Expect to put in more effort than you ever imagined to gain 1 point in 5% of your games otherwise. It's obviously not an efficient way to improve at go.

You have not understood my paper well because

1) it is about increasing one's score by 2 or 4 points - not by 1 point,

2) the summaries alone (at the beginning and the end of the paper) provide already most of the improvement effect while requiring rather little effort - the great effort of learning types and counts by heart can be neglected unless one wants to be perfect,

3) from the paper one can also learn to try to postpone fighting a ko until after the dame - this is easy to learn and nevertheless efficient and relevant significantly more often than 5% (because if one does not understand that the opponent delays winning the ko, then dame ko fight behaviour will occur more frequently in one's games),

4) how long does one need to understand the summaries (a few hours?) - this is more efficient than most other endgame literature with their teaching by scarce example shapes (occurring in 0.005% of one's games or rarer) only approach.

If you want to mention really inefficient study, then why not the mathematical proofs for the dame+ko endgames, which you can find on rec.games.go but which are not in the paper?

Magicwand · Post by **Magicwand** » Thu Jan 27, 2011 12:28 pm

RobertJasiek wrote:You have not understood my paper well because

1) it is about increasing one's score by 2 or 4 points - not by 1 point,

i didnt read that in detail but how is it possible to increase score by 4?
when you have 1/2 point ko all over..it only differ by 1 point max.
from what i read it is a useless math proof (probably not even that more like logic).

RobertJasiek · Post by **RobertJasiek** » Thu Jan 27, 2011 12:40 pm

Recall that the paper assumes Area Scoring! With no dame left, winning versus losing the last ko (all threats being answered because of being bigger than the ko) is a difference of 4 points: The ko is B+2 when filled by Black or is W+2 when filled by White.

Magicwand · Post by **Magicwand** » Thu Jan 27, 2011 1:00 pm

RobertJasiek wrote:Recall that the paper assumes Area Scoring! With no dame left, winning versus losing the last ko (all threats being answered because of being bigger than the ko) is a difference of 4 points: The ko is B+2 when filled by Black or is W+2 when filled by White.

i thought about your logic and it doesnt make sense.
if you are going off area counting and following your logic max is 3 point difference not 4 for situation above.
but i guess if you have 2 step ko and have more ko threat then you will count that as 8 point difference?
then you can make up a example that can make more than 20 point difference according to your logic.

i guess you didnt think about ALL situations before writing your paper.

also.. are you theaching people how to find ko threat?
if you dont have ko threat and he get's final ko then you are not losing anything because it is not your point.
so in my opinion there is no point difference.

RobertJasiek · Post by **RobertJasiek** » Thu Jan 27, 2011 1:16 pm

(It is about area scoring. Area counting is a term used for a different meaning.)

The difference between B+2 and W+2 is 4.

That value is part of the score. Speaking of count, deire value or miai value are other topics. When I say "4 points difference between winning and losing the last ko", then the 4 points refer to the score (like in "Black wins by 30" instead of "Black wins by 26").

With the restrictions stated in the paper's preface, I thought of ALL positions with basic kos, possibly with dame and possibly with ko threats. I have proven ALL related cases. (Part of the restrictions is that the positions shall not have any other endgames nor any teire.)

The paper does not teach how to find ko threats. In fact, "ko threat" in general is still undefined and subject to future research. The paper assumes ko threats in a simplified, abstract and linearly countable sense.

pwaldron · Post by **pwaldron** » Thu Jan 27, 2011 1:36 pm

Magicwand wrote:i didnt read that in detail but how is it possible to increase score by 4?

Perhaps you should read the paper. Robert is quite correct, and has thoughtfully provided examples of such cases.

The analysis is more than just a theoretical curiosity--I have personally seen both the US Open and US Youth Go Championship results decided because of this. In both cases the losing players could have won by playing properly in the dame filling stage.

Magicwand · Post by **Magicwand** » Thu Jan 27, 2011 1:41 pm

OK. you convinced me on that 4.

did you think about this example where black has many ko threat.
black will win all false eye and white have to fill his territory.
that is way more than 4 points.

Click Here To Show Diagram Code: [go]$$Bcm1 Prisoner Count: B-0 W-0 $$ --------------------------------------- $$ | x B B x B x x B x . X . X . X O . . . | $$ | B O B B O B B O X X X X X X X O . . . | $$ | O O O O O O O O O O O O O O O O . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . , . . . . . , . . . . . , . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ | . . . . . . . . . . . . . . . . . . . | $$ ---------------------------------------[/go]

pwaldron · Post by **pwaldron** » Thu Jan 27, 2011 2:14 pm

Magicwand wrote:did you think about this example where black has many ko threat.
black will win all false eye and white have to fill his territory.

Yup, looks like he took care of that too. Second sentence of the paper:

Robert's paper wrote:The positions shall not have teire, any other remaining endgames, nor unusual aspects

Magicwand · Post by **Magicwand** » Thu Jan 27, 2011 3:13 pm

pwaldron wrote:Yup, looks like he took care of that too. Second sentence of the paper:

i guess he took care of that too by deleting it.
ok then what is the main point of the book?

let me sum it up:
last KO can make a maxmim difference of 4 point in area scoreing game.

what other useful infomation is in there other than tedious 1/2 point multiple ko.??

pwaldron · Post by **pwaldron** » Thu Jan 27, 2011 3:39 pm

Magicwand wrote:ok then what is the main point of the book?

let me sum it up:
last KO can make a maxmim difference of 4 point in area scoreing game.

You make it sound like it's a small point. The paper is overly wordy, but several of the examples are worth examining. Players not understanding this strategy has directly affected the result of two US championship-level tournaments.

Life In 19x19

Ko research and theory

Ko research and theory

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?

Re: Choosing ko: How big is too big?