Probe examples

antispin · #1

So – I’ve been trying to get my head round the concept of probes. :scratch:

I’ve read the Sensei’s Library page, and the relevant chapter in Strategic Concepts of Go by Yoshiaki Nagahara.

In the abstract, I understand the idea of a probe as a move that forces the opponent to settle his shape (i.e. choose his strategy), leaving you to adapt your strategy accordingly.

However, I’m not sure whether I actually use probes in my (3 kyu) games. Perhaps I do, but don’t mentally label them as such; or perhaps they are an advanced strategy I haven’t learnt to use yet.

I do make plays which leave an acceptable follow up however the opponent responds. This is surely a fairly basic idea (related to the miai principle). But probes are something more than this, right?

Say I approach a nearly-consolidated territory that has two possible invasion areas. My opponent protects one of the areas, so I invade the other. If that counts as a simple probe, I guess I do use them instinctively.

Or say I invade a framework, having judged that I can live or escape whatever happens. My opponent chooses to seal me in, so I make life while he builds a wall facing the centre. I then jump towards the centre from my neighbouring group, in order to reduce the effectiveness of his influence / patch up a weakness he might exploit / prevent him sealing off a big central territory, etc. Might that sequence be an example of ‘probe’ thinking?

The reason I’m asking is that, to me, it goes without saying that we constantly try to adapt our strategy to counter what our opponent has played. But probes seem to be more mysterious than this… (the professional examples given in Strategic Concepts of Go certainly seem pretty advanced to me, but hell, it’s pro go, what do I expect!)

So, 19x19 people - can you post an example of a probe you played, with an explanation of what you were thinking in terms of possible follow ups? That might help me clarify this concept. Thanks.

HermanHiddema · #2

A probe is a sort of question to the opponent: "Here's a move. You cannot really ignore it, but there are multiple interesting replies here, and which one is the best is not yet clear. Please choose anyway".

The usual follow-up to a probe is to then ignore it, usually for a long time. Example:

Click Here To Show Diagram Code: [go]$$B joseki $$ ----------------- $$ | . . . . . . . . . $$ | . 5 . . . . . . . $$ | a 4 b . . O . c . $$ | . 1 2 O . . . . . $$ | . 3 . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . .[/go]

is a probe.

White cannot ignore it, because then black would play b to capture a stone and take the corner.
There are multiple interesting moves for white. Both a and b look interesting.
Which one is best depends on who gets to play around c first.

White must choose now.

Click Here To Show Diagram Code: [go]$$B connect $$ ----------------- $$ | . . . . . . . . . $$ | . X . . . . . . . $$ | a O 6 . . O . . . $$ | . X O O . . . . . $$ | . X . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . 7 . . . . . . $$ | . . . . . . . . .[/go]

If white chooses :w6:

, black will extend to :b7:

to secure a base for his group.

Later (5 moves, 50 moves, 150 moves, whatever), black can connect at a, which is a large endgame move now.

Click Here To Show Diagram Code: [go]$$B descend $$ ----------------- $$ | . . . . . . . . . $$ | . B . . . . . . . $$ | 6 O a . . O . . . $$ | . X O O . . . . . $$ | . X . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . 7 . . . . . . $$ | . . . . . . . . .[/go]

If white splits with :w6:

, black will also extand to :b7:

to make a base. Black's marked stone is lost for now, but the cutting point at a leaves aji in white's position.

Click Here To Show Diagram Code: [go]$$Bm91 aji $$ ----------------- $$ | . . . . . . . . . $$ | . X . 3 . 7 . . . $$ | O O 4 . 5 O . 1 . $$ | . X O O 6 . . . . $$ | . X . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . X . . . . . . $$ | . . . . . . . . .[/go]

Suppose that many moves later, black plays :b91:

. Can white ignore that? Not really. If white ignores, :b93:

through

completely destroy white's corner. :b93:

exploits the aji of the cut left behind by the descent, which white now needs to defend with :w94:

Click Here To Show Diagram Code: [go]$$Bm91 defense $$ ----------------- $$ | . . . . . . . . . $$ | . X . . . . . . . $$ | O O . . . O 2 1 . $$ | . X O O . . . 3 . $$ | . X . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . X . . . . . . $$ | . . . . . . . . .[/go]

White can defend against the sequence in the corner by exchanging :w92:

for

, but that is definitely an exchange that white would rather not make, so black has still made a profit from the aji in the corner.

In professional play, connecting is more than twice as popular as descending, but both are played regularly.

entropi · #3

Herman, very nice example, thanks. But I am still a bit confused about that specific move :b5:

.

Imagine the same situation where black has not played :b5:

. After let's say 50 moves, black-91 would still force the 92-93 exchange for the same reason, wouldn't it?

If the position remains the same (but without :b5:

), and white ignores black 91, black can then play :b5:

and create the same trouble for white.

On top of that, in case 92-93 exchange later turns out to be uninteresting for black, he could still play the fairly large endgame move at "a" (which would be ruined if white decides to decend).

What am I missing here?

HermanHiddema · #4

Suppose black has not made the exchange in the corner, and black gets :b1:

Click Here To Show Diagram Code: [go]$$B approach $$ ----------------- $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . O . . . O a 1 . $$ | . X O O . . . b . $$ | . X . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . X . . . . . . $$ | . . . . . . . . .[/go]

Does white need to respond with the a-b exchange?

Lets look at what happens if white ignores and black clamps now:

Click Here To Show Diagram Code: [go]$$B clamp $$ ----------------- $$ | . . b . . . . . . $$ | a 1 4 . . . . . . $$ | 3 O 2 . . O . B . $$ | . X O O . . . . . $$ | . X . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . X . . . . . . $$ | . . . . . . . . .[/go]

Compare this to the diagram titled "aji" in my first post. This is much better for white. White's group here is much more stable than before, and black gets fewer point (Note that white can throw in at a then after the capture descend at b in sente later)

But what is black tries to play the same invasion that he did in the earlier diagram?

Click Here To Show Diagram Code: [go]$$B invasion $$ ----------------- $$ | . . . . . . . . . $$ | . 5 8 3 4 . . . . $$ | 7 O 6 . . O . 1 . $$ | . X O O . . . . . $$ | . X . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . X . . . . . . $$ | . . . . . . . . .[/go]

This is exactly the same as the previous clamp diagram, except black exchanged :b3:

for

, which only helps white.

In all cases, white is stronger in the corner, and black's :b1:

stone is weaker in the face of that.

So, the difference is that, given the knowledge of black getting :b1:

(instead of, say, white getting a two point extension there), white no longer has any doubts about the correct response to the clamp. With :b1:

in place connecting is clearly better.

Click Here To Show Diagram Code: [go]$$B bad clamp $$ ----------------- $$ | . . . . . . . . . $$ | . 1 . . . . . . . $$ | 2 O . . . O . . W $$ | . X O O . . . . . $$ | . X . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . X . . . . . . $$ | . . . . . . . . .[/go]

For reference, consider the case where white has gotten the extension :wc:

. Now, if black clamps, the descent is obviously better, because it simply captures the clamp stone. Black clearly loses points here.

So the most important property of a probe is that it forces your opponent to make a choice now, when it is hard, rather than later, when it is easier due to extra information.

cdybeijing · #5

entropi wrote:

Herman, very nice example, thanks. But I am still a bit confused about that specific move :b5:

.

Imagine the same situation where black has not played :b5:

. After let's say 50 moves, black-91 would still force the 92-93 exchange for the same reason, wouldn't it?

If the position remains the same (but without :b5:

), and white ignores black 91, black can then play :b5:

and create the same trouble for white.

On top of that, in case 92-93 exchange later turns out to be uninteresting for black, he could still play the fairly large endgame move at "a" (which would be ruined if white decides to decend).

What am I missing here?

If black makes the probe :b5:

and white responds by descending, that means black can aim at playing 91 in sente. If black plays 91 without first exchanging :b5:

for the descent, then when he later plays :b5:

white will know clearly the best way to answer the move, i.e. by connecting. Black gets a gote endgame move as aji which is likely to be less valuable than playing 91 in sente.

entropi · #6

Amazing! Such a simple looking concept yet so difficult to fully grasp. I understand it now but will I able to apply the principle in my games, that's the question

emeraldemon · #7

just wanted to say that this is a probe I've been trying hard to learn:

Click Here To Show Diagram Code: [go]$$B corner $$ ------------------ $$ | . . . . . . . . . $$ | . . . W . . . . . $$ | . . . X . . . . . $$ | . . . , . . . . . $$ | . . X . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . .[/go]

There's quite a bit of material about it on SL, but I still have a hard time using it in games.

mw42 · #8

emeraldemon wrote:

just wanted to say that this is a probe I've been trying hard to learn:

Click Here To Show Diagram Code: [go]$$B corner $$ ------------------ $$ | . . . . . . . . . $$ | . . . W . . . . . $$ | . . . X . . . . . $$ | . . . , . . . . . $$ | . . X . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . . $$ | . . . . . . . . .[/go]

There's quite a bit of material about it on SL, but I still have a hard time using it in games.

See: viewtopic.php?f=37&t=3511&start=20

My opponent played the well-timed probe at move 37.

Probe examples

Who is online