Depth of reading correlates with rank - and other insights

[HermanHiddema]

Ok, here's a reference position from a real game:

$$W
$$ . . . . . . . . . |
$$ . . . . . X . . . |
$$ . . . . . . . . . |
$$ . . a . . X . . . |
$$ . . O X X W b . . |
$$ . . . O X W . . . |
$$ . . O . O X X . . |
$$ . . . O O . . . . |
$$ . . . . . . . . . |
$$ ------------------+

Click Here To Show Diagram Code

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

This occurred in a game of mine a few years ago at the congress, so there was ample thinking time (2h30m main time each)

This position is the result of a joseki mistake by my opponent, and black has sente. The question is how much aji is left in the marked white stones. Can black tenuki, or would white then devastate the corner?

I spent more than 20 minutes reading every reasonable variation I could find. Some more than 20 moves deep.

Am I sure I considered all variations? No. I was reasonably confident that I had considered most of them. At that point, if my opponent had come up with something brilliant, he deserved to win, IMO.

In the game, I played tenuki, and my opponent made me very happy by exchanging a for b. Because really, there is a lot of aji in the position

Anyway, this is the kind of reading, about 10-20 moves, with maybe some variations a little longer, that is really at the upper end of my ability.

[perceval]

this discussion brings a question to stronger (or not ) player :
do you read the possible answer to a big point ?
what i mean is that i will try to read close contact moves of course, but when i play a big move far from the opponent i never try to read the answer: i just try to get a feel of where the biggest point, but i don't think it involves reading: it's more or less choosing between framework junction or plays in a open area measuring the bigger open area, but it almost always a "static" evaluation so i would say its 0 depth and 3-4 width.

Do you do the same ?

[daniel_the_smith]

this discussion brings a question to stronger (or not ) player :
do you read the possible answer to a big point ?
what i mean is that i will try to read close contact moves of course, but when i play a big move far from the opponent i never try to read the answer: i just try to get a feel of where the biggest point, but i don't think it involves reading: it's more or less choosing between framework junction or plays in a open area measuring the bigger open area, but it almost always a "static" evaluation so i would say its 0 depth and 3-4 width.

Do you do the same ?

I read big moves sometimes. Not sure if more or less than I read other moves. Getting the last big fuseki point is something people my level (should be trying to) do, and that requires something like full-board reading.

[Shaddy]

this discussion brings a question to stronger (or not ) player :
do you read the possible answer to a big point ?
what i mean is that i will try to read close contact moves of course, but when i play a big move far from the opponent i never try to read the answer: i just try to get a feel of where the biggest point, but i don't think it involves reading: it's more or less choosing between framework junction or plays in a open area measuring the bigger open area, but it almost always a "static" evaluation so i would say its 0 depth and 3-4 width.

Do you do the same ?

I read big moves sometimes. Not sure if more or less than I read other moves. Getting the last big fuseki point is something people my level (should be trying to) do, and that requires something like full-board reading.

Uh.. really? I've never even thought about taking the last big fuseki point. Something new every day..

[Knotwilg]

To demonstrate to a beginner that a bulky five can be killed, you may need as many as 13 moves of depth and a width of 2 or 3 at each move. Surely that is not what anybody here means with "13 moves".

Similarly, a pro will not take levels of depth into account that we amateurs still need to read out. They know.
So in many cases, the pro will read less variations and less deeply at that than we (need to). But if pushed to demonstrate, they can, of course.

Neither pro nor mortal needs to take any lessons from a brute force computer there.

Absolute numbers hence don't mean a lot. It's the difference that matters, whether you count up to the elementary killing shape or up to the point of removing the stones.

[daniel_the_smith]

Uh.. really? I've never even thought about taking the last big fuseki point. Something new every day..

Yeah, I saw that on a list of things a pro (can't remember who) tried to teach low dan players. Tedomari isn't just for Bill's endgame problems.

Of course, it's not that necessary if you catch everyone you play in a trick...

[Shaddy]

Hey, I don't play tricks in even games! Usually! Tricks are for scamming low dans out of points in handicap games

[Chew Terr]

Hey, I don't play tricks in even games! Usually! Tricks are for scamming low dans out of points in handicap games

Oooh, so I should learn some tricks, eh?

[daniel_the_smith]

Fortunately I'm not quite strong enough to fall for most of Shaddy's tricks...

[Knotwilg]

With the previous post I wanted to point out the distraction to John's message, caused by discussion the actual absolute count of reading depth. With this one I'd like to go into the message.

It seems like ijime, kokai and "girichon", are trademarks of Korean go, which seems to be all about fighting (down to the wire), counting liberties, bullying and to a lesser extent surrounding (kokai).

The reading of "kokai" was confusing to me: the superficial translation of "surrounding" seemed to emphasize the long term effect of surrounding the opponent, which is felt by the sudden sente falling from the skies in the endgame. But the explanation seemed to point to a more short term, atsumi-like effect.

Interestingly, I've played a game yesterday where I did not bully at all, but instead kept surrounding on a big scale, taking some time for slow moves with big endgame potential. The victory was surprisingly easy, despite the ease with which he could live in my moyo.

Most likely, kokai or ijime understanding will not really matter at my level, but it strangely felt that way exactly.

[John Fairbairn]

Knotwilg: It's kakoi, note. But you are right to point up the fact that kakoi in this case refers to surrounding territory rather than the opponent.

Below is part of a game that illustrates some of the points discussed in the original post. If you want to test yourself, try hard not to read below the diagram, go to move White 56 and try to guess Black 57.

[sgf-full](;FF[3]OH[-B-]AP[MultiGo:4.2.1]SZ[19]
DT[1907]PB[Nakagawa Senji]
BR[6d]PW[Tamura Yasuhisa]WR[7d]KM[0]RE[W+2]US[GoGoD95]MULTIGOGM[1]
;B[qd];W[dd];B[od];W[qp];B[co];W[dq];B[oq];W[cl];B[cf];W[jc];B[fc];W[df];B[dg];W[ce]
;B[ef];W[de];B[fe];W[eg];B[dh];W[fg];B[bf];W[ec];B[id];W[jd];B[ie];W[be];B[ei];W[gh]
;B[el];W[dn];B[do];W[eo];B[cn];W[dm];B[bk];W[bl];B[dp];W[ep];B[cq];W[eq];B[cr];W[ck]
;B[fj];W[je];B[if];W[hi]
(;B[hk];W[jh];B[eb];W[db];B[gb];W[kg];B[da];W[ca];B[ea];W[cb];B[ib];W[lq];B[fn]
;W[fm];B[em];W[en];B[gm];W[pq];B[op];W[pn];B[mo];W[lo];B[ln];W[ko];B[mp];W[jq];B[pl]
)
(;B[jf];W[gk];B[gl]))[/sgf-full]

Before that there is an interesting point about Black 47. Shusai (Tamura) said he expected the variant move shown (call it Black 1), in which case he really had to "make shape on the left by pressing at White 2. But Black would attach at Black 3 and then I can neither block him in nor bully him on the left, and so White might be in trouble." This is an interesting case of ijime (bullying) being applied at the level of high strategy, as the main evaluation criterion almost, and examples like this are why I maintain that the concept has not been properly taken on board in the west.

Back to Black 57. This is me talking now, not Shusai, but I think this is a good example of making three eyes. Black's group is surely alive (two eyes) without this move but would be open to major bullying. Even though there is a vast area on the right where Black would like to get to before White, he deems anti-bullying so important (and, again, note how early in the game) that he makes three eyes to avoid being bullied, giving White sente. Of course Black 57 is not entirely gote as it implies threats against White later on, but still the combination of an apparently small defensive move (a little over 5 points locally?) and some very vague future threats being valued higher than a 20+ point big point is instructive.

[Dusk Eagle]

If black doesn't play , can't white capture black's cutting stone at E14 by playing F14? If black tries to save it, G15 threatens snapback, and if black F16, then G16 seems to kill. I think that's the less abstract reason for playing - it allows black to target a lot of white's weaknesses later on, which he can't do if white has the option available to just capture black's cutting stone.