Life In 19x19

Review: The Theory and Practice of Semeai
Page 1 of 1

Author:  RobertJasiek [ Sat Jul 20, 2013 7:36 am ]
Post subject:  Review: The Theory and Practice of Semeai

Review: The Theory and Practice of Semeai


* Title: The Theory and Practice of Semeai
* Author: Valery Shikshin
* Publisher: Hinoki Press
* Edition: First English Printing, April 2013
* Language: English
* Price: USD 24.50
* Contents: semeai
* ISBN: 978-0-9819121-5-8
* Printing: weak binding
* Layout: slightly inefficient
* Editing: almost weak
* Pages: 219
* Size: 136mm x 212mm
* Diagrams per Page on Average: 1.3
* Method of Teaching: principles, classification, examples
* Read when EGF: 5k - 3d
* Subjective Rank Improvement: +
* Subjective Topic Coverage: o
* Subjective Aims' Achievement: +



Chapters 1 - 4 identify types and cases of semeais and reminds of not wasting ko threats and dissolution near the game end, where appropriate. The first three pages of Chapter 5 and the four pages of Chapter 7 summarise and list the principles and guidelines. The greater part of Chapter 5 and Chapter 6 consist of problems and answers. The book has a nice anecdote in the foreword, a few unrelated notes on Russian go history and finishes with a glossary of Japanese terms and a few photos.


Every semeai type is explained by a few principles (called 'rules'), one or a few examples per principle, a few, or a few more, problems and answers and a summarising list of the principles. The principles pretty much serve their general purpose of identifying a semeai's winner, but do not cover a few cases and sometimes use ambiguous language (for example, "has one or more liberties than" is supposed to mean "has at least one liberty more than" and not "has exactly one liberty or has more liberties than"). For beginners of semeai theory, disambiguation could sometimes be non-obvious.

If a classification into 24 cases by exclusive and fighting liberties were used, 4 cases would be missing entirely and 3 cases can at best be implied due to only a short hint. The book's principles classify slightly differently by using phrases similar to "if the number of liberties are equal to or greater than". Besides other principles, two such principles are used per type. However, to distinguish a necessity of playing locally from a possibility of playing elsewhere, every such principle implies two cases (equally many liberties versus more liberties). In other words, a first impression might suggest that the book's classification needs fewer cases than other schemes, but a closer study reveals that one can express theory differently but not
eliminate cases.

The examples are of the kind 'two basic groups', and mainly for this purpose the principles apply. The examples have shape defects and approach defects. This lets the examples look more interesting, and the reader learns something about counting liberties when there are approach moves, or the semeai formation is not finished yet and there is a fight about the number of inside (called 'shared') liberties. While it is a strength of the book to recommend guidelines for reducing or increasing shared liberties, this new theory in the literature is not explained by generally applicable principles. The generally applicable rules are about the classic liberty counting and semeai winner determination as soon as the semeai configuration is simplified by preliminary moves. So the book offers something noteworthy, but not already a general solution for the phase of unsettled semeai configuations. In this respect, approach moves can occur in the examples, but, for every semeai type and case of approach moves, the reader is left alone with figuring out the implied number of necessary excess and extra approach moves. The author simply presumes the reader to understand these details and other details of similar difficulty. This is one of the reasons why the book is not recommended for players weaker than EGF 5 kyu.

The chapter headings and, where there are some, introductory texts do not always let it be clear which type of semeai is being discussed in a chapter. This, however, is mandatory knowledge for applying a chapter's principles correctly. Here is the decryption: "basic theory of semeai" or "elementary semeai" mean "semeai with 0 or 1 shared liberties and without eyes"; "group without an eye vs. group without an eye" means "semeai with at least 2 shared liberties and without eyes"; "low shape of nakade" means "small eye" (that is, size 1, 2 or 3); "high shapes of nakade" means "big eyes of different sizes" (which, according to the author, can be 4, 5, 6) "or one big eye and one small eye"; "identical high shapes of nakade" means "big eyes of the same size". The phrase "high shapes of nakade" can be confusing, because inside filling stones are ignored and nakade of sizes 4 or 5 exist in different shapes.


When, somewhere in the web, you read "a prepared formula", this is misleading. Although one could speak of formulae for the liberty conditions in the book's principles, it does not have any single formula applicable to all cases of all types of semeais with two basic groups. There is neither the Old Semeai Formula, nor the New Semeai Formula.

The author could not reinvent a general formula yet because of these reasons: his case study is incomplete; in semeais without eyes, he counts all shared liberties of a group to use the condition "has one liberty less" in a rule, instead of first subtracting one liberty; in a semeai with big eyes of the same size, he counts the shared liberties for the second moving player's group. While there are reasons for his kind of study, the combination with the author's apparently missing study of relevant modern semeai theory has the unlucky side effect for the readers, who cannot learn about a general formula from his book.


Instead of using the now pretty established terms "favourite" and "underdog", the author speaks of "strong" and "weak" groups. This choice is a bit unfortunate, because strong and weak are also used in very different meanings to characterise groups. Only within a semeai theory, these terms are clear enough. In a broader context, one might need to speak of a semeai-strong or semeai-weak group. In the book, usually "group" means "string" ("chain").

The book would have benefitted from the term "exclusive liberties". Instead, it speaks of "outside or nakade (eye) liberties".

The author uses "liberties" in three different meanings: 1) "physical liberties", 2) "approach liberties", 3) something closely related to "fighting liberties". However, he does not use (2) and (3) at all. Most of the time, he just writes "liberties", although he occasionally uses the verb "count" in the same sentence when trying to express (3). One might say that he sometimes speaks of "counted liberties" when he means something closely related to "fighting liberties".

The frequent ambiguity of "liberties" makes understanding of the contents harder than necessary. In the same line, liberties are calculated for an earlier position, for a position occurring just after a move during a sequence shown or at the end of a sequence. Typos or thinking mistakes have occasionally led to wrongly stated numbers of liberties.


Altogether, there are 24 problems in the two chapters. About half of the problems, most of which are from games of Russian players, are of intermediate, the other half of advanced difficulty. A few of the examples show important incidents of history. One of them was Ilya Shikshin's last round game resulting in his first European Champion title. I saw parts of that game live and thought: "Interesting!" Unfortunately, the difficult examples have too few variations and this particular game is affected by typos.


Two of the game problems are affected by typos, so that the reader needs to guess the correct positions. A systematic mistake throughout the book affects a few examples, where ko is a possible, but not mentioned alternative to seki or dying unconditionally. The reader must be aware and look for kos not shown in the answers. The English language suffers from too quick proofreading. The author repeats an analysis mistake made also in his other book The Theory and Practice of Analysis.

Although the principles are meant to be valid in general, there are the following problems. Chapter 1 Rule 3 is correct, but alternative play is also possible. "An approach move" in Chapter 1 Rule 4 must be understood as "A necessary approach move", because unnecessary approach moves do not increase the number of approach liberties. Chapter 1 Rule 7 advises to play elsewhere in case of more liberties, but this overlooks the subtle possibility of local negative ko threats.


As the book claims completeness by the definite article in the title, "This book will help you learn how to calculate all types of semeai." in the Introduction and "This book examines all possible types of semeai" on the backcover, the book can be evaluated also by pointing out its missing types of semeais and calculations:

* no general semeai formula for semeais with two basic groups
* incomplete basic case study of semeais with two basic groups
* no general study of semeais with other topologies
* no systematic study of semeais with initially an essential string with a big eye in atari or with unstable eyes
* no systematic study of semeais with multiple groups
* no systematic study of semeais with groups consisting of several strings
* no systematic study of eyes, strong eyes or weak eyes
* no mentioning of nakade of size 7
* no systematic study of semeais with kos (a very complex field!)
* no systematic study of counting approach liberties (basic and CGT)
* no systematic study of shape defects
* no systematic characterisation of liberty types
* no systematic study of endgame values
* no explanation for general validity of the principles
* no algorithms for move orders

Of course, no single book can study everything. Therefore, none should pretend it. With respect to the scope of the book, the first two points combined with improved proofreading would already have sufficed to fit reasonable completeness within the given structure of contents and volume of the book.


Apart from the mentioned mistakes, partial didactic weaknesses, editing and printing weaknesses, missing general formula and slightly incomplete case study of semeais with two basic, ordinary groups, the book is worth reading for learning some basic semeai theory, seeing a couple of interesting problems and getting an idea about early moves leading to a then more easily assessed configuration. One does, however, wish more examples and more variations.

Is the book really suitable as a first book on semeais? The mentioned weaknesses let it be difficult to recommend it without reservation. As a second or third book on the topic, one can appreciate especially the introduction to planning about inside liberties and the, although sparse, entertaining aspects. The principles offer an alternative, but cannot replace complete case studies of basic semeais found elsewhere in the literature.

Author:  jts [ Sat Jul 27, 2013 6:18 pm ]
Post subject:  Re: Review: The Theory and Practice of Semeai

Chapters 1 - 4 identify types and cases of semeais and reminds of not wasting ko threats and dissolution near the game end

Is dissolution a technical term here? To what does it refer?

Author:  RobertJasiek [ Mon Aug 12, 2013 5:06 pm ]
Post subject:  Re: Review: The Theory and Practice of Semeai

"Dissolving a ko" transforms it from being a ko to not being a ko. Means of dissolution can be connecting the ko or capturing some adjacent strings.

Author:  schawipp [ Tue Apr 20, 2021 3:11 pm ]
Post subject:  Re: Review: The Theory and Practice of Semeai

Sorry for unearthing this thread after a long time - I have just dusted off this book from my bookshelf in order to refresh my memory. In chapter 3 (group with eye vs group with no eye) I'm now stumbling across the problem positions No. 4 and 9 (which are actually the same positions just with black to move in the first case and white in the second).

There is a white group with 6 effective outside liberties plus an eye sized four points with already one black stone in. Apparently the eye is counted as 5 nakade liberties minus the one stone; thus 4 eye liberties.

My issue is that black can play atari inside of the eye, thereby reducing it to a three-point eye with two black stones in (if w replies). That would be only one eye liberty left instead of four, and then the problem becomes somewhat pointless.

If someone who also has this book could check these problems in chapter 3 (and either confirm the mistakes or correct me) that would be great.

Author:  Bill Spight [ Wed Apr 21, 2021 10:11 am ]
Post subject:  Re: Review: The Theory and Practice of Semeai

RobertJasiek wrote:
"Dissolving a ko" transforms it from being a ko to not being a ko. Means of dissolution can be connecting the ko or capturing some adjacent strings.

Some native speakers talk of resolving a ko. :)

Page 1 of 1 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group