Tesuji are group of skillful moves, but can it be said that these are moves that are universally good? When given the chance, play tesuji? Which tesuji is appropriate depends on stage of game and value of groups, of course.

Take for example the snapback and connect-and-die tesuji:

Click Here To Show Diagram Code

[go]$$B
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | X X X . . . . . .
$$ | O O X . X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X Q Q X .
$$ | . . . . O a b . .
$$ ------------------[/go]

Marked stones will die if black plays here. Both "a" and "b" work, but "a" is a snapback. Should black play it because it's tesuji (assuming it's that stage in the game where these stones are valuable enough)?

This position conveniently has another application:

Click Here To Show Diagram Code

[go]$$B
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | X X X . . . . . .
$$ | O O X . X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X O O X .
$$ | . . . 2 O 1 a b .
$$ ------------------[/go]

Black wants to make snapback and plays 1. White responds with solid connection 2. Should black take the stones with "a" or create snapback with "b"? This is becoming closer to the kind of universal rule I'm looking for. If it doesn't really matter which one you play to achieve the primary objective, would tesuji play always create better position?

How about defense? Is the best kind of defense the type of play that prevents tesuji?

Click Here To Show Diagram Code

[go]$$W
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | X X X . . . . . .
$$ | O O X . X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X O O X .
$$ | . . . . O 1 . . .
$$ ------------------[/go]

There's also third example I'm curious about shortage of liberties tesuji. This position can happen commonly in endgame:

Click Here To Show Diagram Code

[go]$$W
$$ | . . . . . . . O .
$$ | . . O O O . . . .
$$ | . . . . . . O . .
$$ | X X X 1 b . . . .
$$ | O O X a X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X O O X .
$$ | . . . . O O . . .
$$ ------------------[/go]

White peeps at the cutting point. Black can connect at "a" or create a simple shortage of liberties situation with "b". As far as shape goes, would it be better to utilize the tesuji nature of "b"? As for white's initial plan for peeping, considering it's countered by this tesuji, would it be better to play "b" instead to achieve the same eye-stealing effect?

If you care about winning, I wouldn't get hung up on if a sequence seems like a tesuji. Try to play the best move. Given two choices, one might be better due to ko threats, for example. But it makes no difference, I think, if it's a "tesuji".

If you want to look cool, always play the tesuji, though

_________________
be immersed

look cool will kami no itte be a tesuji move?

_________________
长考出臭棋.

The questions you pose seem a bit odd to me. To me, a tesuji has to satisfy two conditions:

1. It is the best move that satisfies an ultimate purpose.
2. It is skillful/clever/beautiful.

Take the following board position as an example:

Click Here To Show Diagram Code

[go]$$Bc
$$ ---------------------------------------
$$ | . . . . . . . . . . . . X X X O . W X |
$$ | . . . . . . . . . . O . X O O O O X X |
$$ | . . . . . . . . . . . . X X X O O O . |
$$ | . . . O . . . . . , O . . . O X X O X |
$$ | . . . . . . . . . . . . O . O O X O X |
$$ | . . X . . . . . . . . . . . . O X X X |
$$ | . . . . . . . . . . . . . . . O O O . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 1 . . . . . . . . . . . . . . . |
$$ | . O . . . O . . . . . . . . . . . . . |
$$ | . O X X X O . . . , . . . . . X . . . |
$$ | . O O O O O . . . . . . . O . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Should White play the "tesuji" of D5 here or play the simple capture of T17? Even on an empty board with nothing but the bottom left stones, D5 would not be a tesuji, and in this position it clearly is not a tesuji. It is only a tesuji when the ultimate purpose that needs to be achieved is to kill those 3 Black stones as White, which is not the case here. So should T17 be considered a tesuji because it's the best move? No, because it is not clever or skillful.

If we define a tesuji this way, it immediately answers all of your questions: between 'a' and 'b' in your examples, choose the one that's higher in endgame value. If it's the best move on the whole board, then it may be considered tesuji if it is clever/skillful. What appears to be a move of cleverness and skillfulness is all relative; to a 30k the snapback may appear very clever; but to a 9p it is just as plain as any other move.

To me a tesuji is just some abstraction which helps you read. It's only purpose is to make you easily spot a sequence you wouldn't see that quickly(or at all) without knowledge of the tesuji.

_________________
Sanity is for the weak.

I'd second Araban's (to me; rather obvious) comment, with one change; rather than just say "must be clever" I would add tesuji are defined by

>Not obvious without sufficient reading

That's what's always defined tesuji to me. Albeit, they're often found by recognising starting shapes without (re)doing the reading.

_________________
Revisiting Go - Study Journal
My Programming Blog - About the evolution of my go bot.

Since it's only a matter of definition, the discussion is likely to be fruitless but I still want to say that to my understanding tesuji is just a local concept.
It is obvious that Araban's example (move D5) for capturing the anyway useless stones is a bad move at that stage of the game. But according to my interpretation, a move does not need to be a "globally" good move in order to be qualified as tesuji. D5 is still a tesuji but a badly (bad timing) played tesuji. Maybe at a later stage of the game it will become the biggest move on boards. But again, this is just a matter of definition.

_________________
If you say no, Elwood and I will come here for breakfast, lunch, and dinner every day of the week.

I too see tesuji as local tactics. Hence it may not be the biggest move on the board. Consider a 4pt tesuji vs a 10pt side extension. Obviously you choose the side extension even if nobody would call it a tesuji.

_________________
FusekiLibrary, an opening library.
SGF converter tools: Wbaduk NGF to SGF | 440 go problems | Fuseki made easy | Tesuji made easy | Elementary training & Dan level testing | Dan Tutor Shortcut To Dan

My personal definition of tesuji is a move which seems/feels like some kind of magic, the best move in a local fight which isnt easy to find if you dont know that it is there. So I think which move is a tesuji and which not differs from rank to rank. For me a move like this 'a':

Click Here To Show Diagram Code

[go]$$B
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | X X X . . . . . .
$$ | O O X . X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X Q Q X .
$$ | . . . . O a b . .
$$ ------------------[/go]

is certainly not a tesuji.



You should not play a move because it is a tesuji but because it is the best move in the situation. Sometimes this move feels so brilliant and magical that we can call it tesuji
A Tesuji which results in a bad position cant be a tesuji per definition.

- That's what I meant. There's consensus that some moves are tesuji. Many authors call snapback a tesuji. As they are tesuji, they are by definition the best moves to satisfy certain kind of purpose. I wonder if this has been verified by experience.



- Cleverness and skillfulness are pretty vague. I don't think it's wise to disregard elementary tesuji just because everyone knows about them. If I were to build a strong Go bot, should I give higher priority to tesuji type moves locally? Here "a", "b" and "b" for diagrams in opening post.

The point is to play the best move. This may or may not be a tesuji.

_________________
be immersed

I think that "tesuji" just means "the right move". It has got a connotation of that move being not quite obvious in western literature, but I am not sure if this is actually a good definition (define "obvious"...). It is perhaps best thought of as "the right move in this shape".

_________________
A good system naturally covers all corner cases without further effort.

- This truism won't help in actually finding the best move. People read tesuji books because they want to improve their game. Which move would you play in examples listed in OP?

No, "honte" is the right move. Tesuji is a move that is a bit of a "trick" in some ways. I think the translation "skillful finesse," while inelegant, does sum it up well.

_________________
My blog about Macs and more: Kirkville

From Fujisawa Shuko, Dictionary of Basic Tesuji Volume I: Tesuji for Attacking:

"Tesuji is the name for moves that make the most effective use of stones... Until recently, common usage has been that the term 'tesuji' by itself refers to techniques for close-combat fighting; 'suji' refers to key points for attacking and 'shape' to key points for defense. However, it is also the fact that people talk about 'tesuji for life and death' and 'yose (endgame) tesuji', etc., so among these various usages the definition for tesuji has become a bit fuzzy....

In the midst of this widening use of the word, I've attempted to categorize moves by their purpose. It could be said that any move that's not a bad move is a tesuji. It can also be said that whether or not a move is bad depends on the whole board situation. Of course, there are many debatable features to the categorization of this book..."

I don't think there exists a "silver bullet", like you seek. The answer is always to read and evaluate.

Here are some ideas of how I might think:

Click Here To Show Diagram Code

[go]$$B 
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | X X X . . . . . .
$$ | O O X . X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X Q Q X .
$$ | . . . c O a b . .
$$ ------------------[/go]

Above, let's say black plays "a" and white plays "c".
Black may decide to tenuki, but if he does and white later captures the black stone, white gains a point. If black plays "b" immediately, and white does "c", if black does tenuki, white playing "a" doesn't give an extra point.

On the other hand, if black plays "b", white may be less inclined to even answer than if black plays "a".

So instead of trying to develop an always working heuristic, I think it's better to consider the implications of each move, given the entire board.

Click Here To Show Diagram Code

[go]$$B 
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | X X X . . . . . .
$$ | O O X . X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X O O X .
$$ | . . . 2 O 1 a b .
$$ ------------------[/go]

Both have goods and bads. Suppose you want to preserve ko threats. Taking the stones off the board removes the ko threat white has of saving the stones. I this sense, "a" may be better.

But suppose that there is a weak black group to the right of this, and playing "b" will allow you to connect. In that sense, "b" could be better.

Click Here To Show Diagram Code

[go]$$B
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | X X X . . . . . .
$$ | O O X . X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X O O X .
$$ | . . . c O 1 b d .
$$ ------------------[/go]

As white, I'd usually connect at "c". If I play "b", black gets "d" for free, and I still have to connect at "c" after he captures, anyway...

BUT, let's say the position was ignored for awhile, and now it is ko somewhere on the board. Playing "b" gives me more ko threats.

Click Here To Show Diagram Code

[go]$$W 
$$ | . . . . . . . O .
$$ | . . O O O . . . .
$$ | . . . . . . O . .
$$ | X X X 1 b . . . .
$$ | O O X a X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X O O X .
$$ | . . . . O O . . .
$$ ------------------[/go]

"a" could be good for giving white less ko threats, and "b" could be good for taking away white eyespace, for example...


So basically, go is too complicated to have always working heuristics like this. That's why reading and evaluation are important.

It is great to study tesuji - it gives you great ideas on possibilitiea you could do... But you should still play the move you think works best with the current board position, whether or not it is in a tesuji book.

_________________
be immersed

I have always thought of tesuji as local moves, except if kos or ladders are involved. It is always used to achieve something local, e.g. connect/save own stones, cut/capture the opponent's stones. And the type of move deviates from trivial/shape moves, so it's a move that has to be learned beside the "normal" moves.

Click Here To Show Diagram Code

[go]$$B
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | . . . . . . . . .
$$ | X X X . . . . . .
$$ | O O X . X . . . .
$$ | . . O X . . . . .
$$ | O O . O X X X . .
$$ | . . . O X Q Q X 6
$$ | . . . 2 O 4 1 . .
$$ ------------------[/go]

If Black plays 'b' then White 4 is almost sente, as it threatens the attachment at 6 (or cutting). I think this fact is worth more than the potential single point in most positions.

It could very well be, and it is a good aspect to consider. My point is just that relying upon whether something is considered as a "classic tesuji" or not is not sufficient for playing the best move. Considering possible consequences of a particular move, like you've done here, is important for having a good strategy.

_________________
be immersed

Which move perceived as a tesuji depends on a player's strength. To 30k, Crane's nest is impressive, but to a dan player, it's an ordinary move that they know by heart.