Review
Robert Jasiek, Capturing Races 1 - Two Basic Groups, English, Berlin 2011, 272 pages, 28.00 Euro (PDF 14.00 Euro).

Reviewer: Thomas Redecker
Originally published in Deutsche Go-Zeitung (= ‘German Go Journal’), issue 1/2012, p. 22-23


Introduction

Anyone who wishes to know everything -- and I mean really ‘EVERYTHING’ -- about the supposedly ‘trivial’ topic of a fight between only one solid string of stones of both parties is well served by reading this book.

In order to first guard interested readers against potential misunderstandings, let me point out that the book deals with the “IF” -- not with the “HOW”.

The book covers the fundamentals for answering the question “Should I play here?”, and sensitises the reader for several relevant cases. E.g. for a group that cannot save itself any more, and that becomes, unnecessarily, killed for a second time, just to be on the safe side. Or a Semeai, in which one side is so much behind that any attempt to rescue would mean only wasting a move. In the future, the reader can concentrate his/her efforts on capturing races in which he/she can gain something. I.e. saving an own group that the opponent cannot capture any more, or obtaining a group which otherwise could have been rescued by the opponent.

The book does not provide ‘techniques’ in the strict sense. I.e., the reader will not find the treatment of Tesuji of the ‘Life&Death’ kind because the book really is not about HOW best to kill an opposing group -- but ‘only’ about IF one was to strike again.

As can be expected from the author, the presentation of the book’s contents is particularly comprehensive. Used terms, definitions, concepts, etc., are consistent, and explained well. The English is easy to read.


Contents

Chapter 2 lays the foundations for the theory building that will be constructed during the book, and specifies -- in great detail -- the later used terms that are partly newly introduced by the author. One or the other reader will be familiar with terms such as ‘favourite’ or ‘underdog’ from Richard Hunter’s book “Counting Liberties and Winning Capturing Races”. However, the reader should be prepared that -- in comparison -- the presentation in the reviewed book is considerably more extensive and exhaustive.

In great detail, chapter 3 treats, as ‘Class’ 1, the parallel arrangement of two basic groups. As subdivision, the author introduces 5 types with up to 6 cases, which are partially re-used, partially extended in the following chapters.

Chapter 4 presents a new Semeai formula that is elucidated and elaborated in great detail -- for all types and cases previously introduced for Class 1.

Chapter 5 introduces, as Class 2, the ring formation (one group surrounds the other) of two basic groups.

Chapters 6 to 9 describe, as Classes 3 to 6, one or two groups (in parallel arrangement) in Atari, or with open Nakade. These chapters are considerably more compact than those for the previous classes.

In chapter 10, the author deals with earlier findings from the literature -- these are compared with his own insights, and offers corrections for identified weaknesses.

A bibliography and a subject index complete the book.


Presentation of the Material

In my estimation, the book’s content is refined in such a way that each type of reader will find sufficient material to approach the issue presented. The usage of many diagrams -- for extensive, and detailed, explanations -- is helpful, so that also such readers can have a very good time who are less familiar with mathematics or computer science.

Each chapter starts with a couple of problems that the reader can deal with -- however, does not necessarily need. There are always compilations of key messages, summaries of the previously written, repetitions of essential contents at the sections’ or chapters’ ends, as well as a somewhat formalised tabular short summary, and finally the solutions of the problems from the chapter’s start.

The chapters’ structure to teach the Semeai classes is uniform. The underlying structure of sub-chapters as well as the always homogeneous compilation and structuring of the actual content secures a high recognition value. For this reason, and because of the very comprehensive treatment, for me, the book has also characteristics of a dictionary, or of an encyclopaedia.

A certain formalisation -- typical for the author -- can be noticed but is not that extensive to be distracting. Only occasionally the author loses himself in the danger of wanting to be all too exhaustive, and to not miss any aspect.

In my opinion, chapters 6 to 9 (Classes 3 to 6) are out of the ordinary, compared to the previously explained Classes 1 and 2 -- especially related to the evaluation of practical applicability. Obviously they are required for a complete description, also in the view of the planned subsequent volumes.


Layout

With respect to layout and used fonts, the book is an easy read. I consider it a convincing idea to use a bigger font, combined with a more generous text arrangement, for the theory sections than for those parts of the books that comment examples in detail.

In my opinion, the diagrams are unusually large in relation to the used font size; the capital letters therein are too massive. Presumably this is due to the parallel supply of a screen edition.


Missed Chances

My feeling is that the reader is not sufficiently guided in one decisive aspect -- and this is the necessary paradigm shift required for the study of Semeai, for those who come from the world of ‘Life&Death’ / ‘Tsume-Go’. Contrary to the usual experience, two eyes are absolutely counter-productive when dealing with ‘Semeai’. If one side has got the benefit of two eyes, the ‘Semeai’ issue is grounded from the outset. In the same sense, it is already harmful to have the potential for getting two eyes.

The tables would have profited from a further systematic compression, for the sake of better recognition of overall dependencies by the reader.


Usefulness

To really gain from the book, it will be necessary to read it several times. Even then, after finishing the book, the one or the other reader will stay behind with an unspecific feeling of having missed something decisive. However, this is not true. You have to hold in mind that the book treats only comparatively simple Semeai. All the difficulties that you will encounter in real life are reserved for subsequent volumes.

So, I cannot answer the question to what extent reading of only this volume of the series helps for strength improvement.

Regardless, the book provides an important step forward in the field of go theory.

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Thanks for your thorough review.
You mean striving for two eyes reduces your changes to kill the enemy?

Semeais can have eyes but each group in a semeai cannot have two or more eyes because then it would not be a semeai any longer but independent life. Therefore, simply speaking, a "semeai eye" is something not partitioned into two regions (yet) while an "independent life eye" is one of already at least two separate regions.

Killing a dead group or bringing to life a living group is wasting one move, essentially equal to passing and so (early during a game) about equal to almost one handicap stone or rank. Capturing races do not occur in every game, so getting sente related to death and life in semeais right is, on average worth less. If a semeai occurs once every, say, 3 games, then such a mistake equals almost 1/3 of a rank.

Everybody has some preliminary knowledge on semeais and so the book does not substitute totally absent knowledge by complete knowledge. Rather the book can substitute partial knowledge by complete knowledge about the question of taking sente (beginners: several times taking sente) in case of the basic semeais explained in the book. So to estimate the book's effect on a player, the fraction of his prior partial knowledge would have to be known.

Another fraction is relevant: the restriction of Volume 1 on basic capturing races and the implication of that knowledge for more complicated semeais. Some of them are decided by the book's knowledge, others are decided by the additional aspects of extra complications: approach defects, shape defects, kos etc.

Hence you are right that it remains difficult to estimate a reader's likely rank improvement due to the book. The conveyed knowledge is essential but partial. Similarly knowing all tesujis but not being able to identify a semeai's type would be a different part of essential but partial knowledge. It is best to have knowledge of both fields: semeai types and tesuji.

More generally, it is best to have preferably complete knowledge of all fields of go theory:)

My confusion is not completely solved yet ( will it ever ? )

To me it seems that even if you have the potential for two eyes you might decide to fight out the semeai in order to kill the enemy group. In the proces you might loose your potential but you calculate that it will be restored after the kill.
As a side path, it seems to me that the potential to win a semeai most of the time ( i.e. apart from nakades ) implies the potential for two eyes.

Therefore the book solves the classes 3 to 6, which initially have at least one group in atari or at least one unstable big eye (a semeai-eye which threatens partition into two eyes).



Your confusion still exists because you still do not distinguish
- life-and-death-eyes in settled positions (can be transformed to become part of a two-eye-formation or, in case of a seki, is surrounded by uncapturable strings),
- life-and-death-eyes during the middle game,
- combinatorial game numbers a la Landman,
- nakade shapes (which can occur in semeais or elsewhere),
- surrounded shape in semeais (for which the book defines semeai-eye, i.e. those things you call an "eye" but occurring in some semeai group).

In particular you seem to be confusing life-and-death-eyes with semeai-eyes.

When a semeai is dissolved (one side has been killed), then the opponent has lots of life-and-death-eyes or their eyespace potential there. As long as the semeai is still there, there can but need not be semeai-eyes but usually most parts of a semeai do not consist of them.

Dear cyclops,

Let me try to explain, although this specific issue is not trivial at all, and might not be easily understood.
It took me 1.5 times of reading the book to become aware of the effect. Thereafter, I did not encounter any difficulties any more to understand its contents.

“Semeai” (攻め合い) is a combination of
semeru (攻める) = to attack / au (合う) = to fit, to join
So “semeai” means something like “attack together” / “mutual attack” (hope this makes sense, my English is not the best).

It’s similar to “miai” (見合い).
miru (見る) = to view / au (合う) = to fit, to join
So “miai” means something like “view together” / “consider together”

Semeai is NO synonym for “Life & Death”, which represents the biggest part of “Tsume-Go”.

The usual Tsume-Go theme is of the “Black to kill / White to live” kind. However, the description of the aims is not exhaustive at all.

It should be “Black to kill White’s group, assuming that the surrounding black stones are alive.” / “White to live, at least with the main part of her stones, inside Black’s stones, which are assumed to be alive (with the exception of some stones that might be captured during the living sequence)”.

This means either “Black to prevent two white eyes” or “White to get two eyes”.

Let’s limit ourselves now to White’s side of the medal, to simplify the following text. Black’s side is complementary.

In general, White’s aim can be achieved by just occupying the vital point of the problem. That’s all.
There is no need to count liberties, approach-move liberties (, and Ko-shapes).

After White obtained two eyes, there is nothing Black can further “attack”. This means that there is no “Semeai” any more. Getting two eyes is the very simple, but minor, part of the whole story.

It is by far more complicated to calculate positions, in which groups of both parties “fight for Life & Death” that cannot get two eyes. If we consider “Seki” (= “mutual life”) to be a special case, this results in a fight to the very bitter end. Both sides must attack, attack, attack.

Counting liberties becomes very important now, counting approach-moves, too, and also counting Ko-shapes, and the order of Ko-captures. This is the by far more complicated part of the topic.

Thomas

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Strictly speaking, Black could try (multi-step) ko threats to resurrect the semeai;)

Such misses the point.

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Thx, you guys, for your extensive replies. I will chew on it for some time. Probably it is best to read the book to fight some remaining confusion.