Hi everyone,

Looking for advice on the appropriate level to read "The endgame" and "Get strong at the Endgame". I feel like I've got little to no idea what to do during the endgame aside from looking for "big-ish" "sente" points and playing them. No ideas really about endgame tesuji or counting or anything.

Is it possible to work through 'Get Strong At' without having read the Endgame book? I find problems a lot of fun and easy to fit into my daily life; but theory books can be a little dense and sometimes feel like a chore to read..

Should I be worried about this at KGS 8k? Or should I just stick to my current program of tesuji/l&d problems, playing and reviewing?

Hi Zac, that's a very good question. (And a very popular one here and on other servers, too.)
There's another recent thread, I think by Kirby -- study what you enjoy.
Pros have to work very hard because it's their livelihood; they have to feed their family.
For us amateurs, it's different. If it's a big chore for you, you're not going to stick to it anyway,
not for long. So, find some method(s) that you enjoy and that you can sustain over a longer period of time.

I don't own "The Endgame" but I got "Get Strong at the Endgame" through SmartGo Books. I'm maybe a little stronger than you, not enough to make any difference I should think.

The easy question first: Do you need this? No. You can ignore this beyond the simplest rules of thumb since games at our level are more often decided long before the endgame begins rather than in the endgame so the extra 3 or 4 points you might pick up here won't really help cover the 20 point errors you're making in the midgame already often enough to make the time spent doing this instead of working explicitly on your reading worth it.

The second question: You don't need to read "The Endgame" first similar to how you don't need to read "Tesuji" before doing "Get Strong at Tesuji." Would it help? Probably. Would "Get Strong at..." help reinforce what's in "The Endgame," almost certainly. So combining them is an advantage but not absolutely necessary.


I picked up GSATE purely out of curiosity and I dip into it now and then because I enjoy endgame problems. I also do endgame problems on GoChild. In the GoChild ones you're still reading sequences out, you're just not looking to kill or live so it's still reading practice. GSATE is a combination of counting and looking for tesuji. Fun, I'm not sure how relevant it is to weak SDKs though.

As far as endgame plays go, you are concerned with the evaluation of plays, with sente and gote, with the sequence of play, and with tesuji. If plays are simple, then evaluation and determining whether a play is sente or gote or ambiguous are easy. The question of the sequence of play is not so easy, because it is concerned with the choice between plays of the same size, or with exceptional cases: with making a smaller play before a larger play, or with ignoring a sente, or with playing a sente before the opponent will answer it, and so on. Endgame tesuji have varying degrees of difficulty.

You do not have to be very strong to learn the basics of evaluation and classification of plays in terms of sente and gote. Just start with simple plays. Once plays are no longer simple, then the difficulty of those tasks becomes much greater. As a practical matter, you have to know which plays to consider and which plays to ignore. And to do that, you have to know tesuji.

As for studying the sequence of play, if you can't evaluate and classify plays, you cannot even start. That is an advanced topic. Most problems in the sequence of play that you find in books or magazines simply require you to make the plays in order of size.

Some years ago I considered writing an introductory endgame book, aimed at 10 kyu players. After looking at a number of 10 kyu vs. 10 kyu games, I realized that the main thing that 10 kyus needed to learn about the endgame was how to handle shortage of liberties. As liberties get filled at the end of the game, opportunities to exploit the shortage of liberties arise. Big swings can occur. The ability to distinguish between a move than gains 2 points and one that gains 1/2 point is not as valuable as the ability to exploit a shortage of liberties or to protect against such exploits.

You are more advanced, but I suspect that the most practical study of the endgame for you would be endgame tesuji and shortage of liberties. Besides, you can quickly learn the basics of evaluation and classification.

Edit: Since you are interested in the endgame, do review your games backing up from the end. There is no reason for an SDK not to play the last moves on the board perfectly or nearly perfectly. Furthermore, you can usually work out best play yourself.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

I don't know the book Get Strong at the Endgame. The book The Endgame, as I recall, is mostly about sente/gote and evaluating the size of moves. Another book that might be useful is Slate and Shell's 200 Endgame Problems by Shirae. It is almost entirely about endgame tesuji. There are many "standard" endgame tesuji which occur often. They are very useful to know because there are Knee-jerk type moves which amateurs often play when a standard tesuji could get them quite a few more points. 200 Endgame Problems is written for kyu players.

The counting part is easy to fix. Just do it. Counting is one of those things most people know how to do even without playing Go =). In the endgame, you can often read right to the finalization of borders, so there's much less of the 'guess where borders will be drawn' involved in middle game counting. Read out where the borders are formed if you play first, if your opponent plays first and see what the territory difference is between the two. Now you know which gote move to take. Ignoring a sente-ish move can be harder to count, but if you can at least estimate what damage you'd take, you can play a bigger sente move in response to your opponent's move. The challenge is just in having the patience to do all the counting, but there's no reason you can't make correct moves fairly close to the end.

The conventional wisdom is that games our rank aren't usually decided within the margins of good endgame play, but I've certainly won and lost games by it. It can also be a confidence boost: knowing that you've got a better endgame than most players your rank can allow you to play calmly from a slight disadvantage rather than feeling forced into a desperate invasion.

I think that an endgame counting problem would look something like this.

Click Here To Show Diagram Code

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

How much territory does Black have in the corner?

(Hat Tip to lightvector. )

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

_________________
"The more we think we know about
The greater the unknown"
Words by neil peart, music by geddy lee and alex lifeson

Thank you all for your responses.

I guess I find the topic interesting, so, what's the harm in studying it if I like? I might leave 'The endgame' until later and consider picking up both GSA and 200 Endgame problems. Practicing counting certainly can't hurt, especially for being able to clearly visualise what the board will look like if I take a certain move or if my opponent takes a certain move. The tesuji will certainly help and will give me more much needed reading practice.

On a related note, I've started watching a few lectures from internetgoschool.com, and in the step by step course Guo gives a counting exercise at the end of the lecture.

As to the problem Bill posted;

Bill's question:

Ah, so that's how one comes up with brilliant tesujis like black 1 in the first diagram.

Good thing I don't play when I'm that tired

EDIT: Don't know what I was thinking.... see my later post for a more complete answer

Click Here To Show Diagram Code

[go]$$Wc White sente
$$ +--------------
$$ | . . 1 1 O . .
$$ | X X . X O . .
$$ | O . X O O . .
$$ | . X O . . . .
$$ | 1 X O . O . .
$$ | . O . . . . .
$$ | . O . . . . .
$$ | . . . . . . .[/go]

White has three sente plays.

Click Here To Show Diagram Code

[go]$$Wc
$$ +--------------
$$ | . 2 1 . O . .
$$ | X X . X O . .
$$ | O 4 X O O . .
$$ | . X O . . . .
$$ | 3 X O . O . .
$$ | . O . . . . .
$$ | . O . . . . .
$$ | . . . . . . .[/go]

is the least complicated way. leaves the two way ko threat involving .

Click Here To Show Diagram Code

[go]$$Wc
$$ +--------------
$$ | . 4 1 3 O . .
$$ | X X 2 X O . .
$$ | O 6 X O O . .
$$ | . X O . . . .
$$ | 5 X O . O . .
$$ | . O . . . . .
$$ | . O . . . . .
$$ | . . . . . . .[/go]

is another way.

Click Here To Show Diagram Code

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

gives Black the chance to make this blunder. Why is it a blunder?

Click Here To Show Diagram Code

[go]$$Wc Normally a mistake
$$ +--------------
$$ | . . 4 3 O . .
$$ | X X . X O . .
$$ | O 2 X O O . .
$$ | . X O . . . .
$$ | 1 X O . O . .
$$ | . O . . . . .
$$ | . O . . . . .
$$ | . . . . . . .[/go]

gives Black the chance to make ko, if he has enough good enough ko threats. White can avoid the ko by jumping to 4. Assuming that Black can win the ko, who has normally made a mistake?

Click Here To Show Diagram Code

[go]$$Bc Reverse sente
$$ +--------------
$$ | . . . 1 O . .
$$ | X X . X O . .
$$ | O . X O O . .
$$ | . X O . . . .
$$ | 1 X O . O . .
$$ | . O . . . . .
$$ | . O . . . . .
$$ | . . . . . . .[/go]

Black has two reverse sente plays. The one on the left makes the standard assumption that the surrounding region is White territory. How much does the reverse sente gain? Which play is better in terms of ko threats? If ko threats do not matter, which play is better?

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

reverse sente for black will not happen because white has right to the endgame.
it is meaningless to talk about what happens when black plays it first.
that is why there is only one answer which is 3 point for black.

_________________
"The more we think we know about
The greater the unknown"
Words by neil peart, music by geddy lee and alex lifeson

The usual case is to assume White's privileges. Under exceptional global circumstances (such as currently Black to move and only tiny endgames available elsewhere), Black's reverse sente can occur. Therefore, in the general case, the set of solutions contains more than one answer.

Doing anything when you'r that tired isn't a good idea.

in 19x19 board you are saying that white doesnt play that privileges until only tiny endgames available elsewhere?
lets stick to what is real. current question is asking what is the value of the black territory.
if you say anything other than 3 then you so dumb that you dont understand the point of the question.

_________________
"The more we think we know about
The greater the unknown"
Words by neil peart, music by geddy lee and alex lifeson

I don't know if reading those books are right at this stage, but I agree with those who've said attention to the endgame is worth it at this stage. If trying to get stronger is a thing you want to do (and it doesn't have to be to enjoy the game) then it is definitely worth it. It is a thing of beauty in its own right. And if you want to win more games, it helps.

About getting stronger: Assessing reasonably your endgame moves is easily doable at 8k. The branches of the tree are so much smaller and so much more assessable that you can do a basically good job without a lot of trouble. Sure there are complexities and subtleties that are way too much, but that has no bearing on whether it is easy to learn the basics. The real kicker to me is that all the skills you use to manage the endgame ... counting the board, assessing sente and gote, counting the score difference between pairs of positions, reading out small tesujis ... are skills you can use throughout most of the game but are most easily applied in the endgame. Carefully apply these four kinds of skill in the endgame and it gets easier and more natural to apply them through-out; your knowledge and facility can grow from the endgame back through the rest of the game. One can't solve go completely, but the concept of solving by backward induction still applies. Endgame can also be good for mental toughness and fighting spirit.

A different sort of argument, and I suppose a subjective one, is that the patterns and skills applied in the endgame are an elegant landscape in their own right. Maybe not everyone's first choice of beauty, but still beautiful. (Not everyone would choose to study calculus, but if you have to as part of your other studies, try to enjoy it. It is one of the astonishing achievements of the mind and beautiful in its own right.)

At a slightly grubbier level, you can win games at your level by knowing good endgame. Lots of them. For some reason, not many people agree with me that endgame is beautiful and they don't study it or put energy into it in kyu grades. The result for me was that until I got to about 2k, if I was less than 15 points behind in late middle game, I could win a lot more often than I lost. Put a different way ... an 8k with really good endgame _is_ a 5k. A bunch of people have said that the score differences are so much bigger at 8k that endgame isn't a factor yet. True about the dispersion, but not about the factor. Theoretically, perfect endgame between two strong players produces no difference; they both identify who gets that last point and make sure that is the result. You're not playing against them. You're playing against people who don't know quite what they're doing and don't count much. You can get a 15-point swing from their mistakes. Even with a wide dispersion of scores, moving the bell curve 15 points in your favor is good.

I agree with most of what you say - endgame is certainly important. But to me, an 8k is... an 8k. If his endgame is good enough to have a winning rate equivalent to 5k, then he's a 5k.

_________________
be immersed