(;GM[1]FF[3]
VieW[]SZ[19]
PB[Kitani Minoru]
BR[9 dan]
PW[Go Seigen]
WR[9 dan]
GN[1st Japan's Strongest Deciding Matches]
EV[Go Seigen vs Kitani Minoru]
DT[April 14-15, 1957]
PC[Tokyo, Japan]TM[10 hours each]
RE[W+Resign]
SO["Detailed Analysis of Wu's Famous Games. Vol.3" pp. 95-188. People's Sports \[China\], 1986]
KM[0]
C[Go Seigen (Wu) vs. Kitani

(The 1st Japan's Strongest Deciding Matches)

Black: Kitani Minoru, 9d

White: Go Seigen (Wu), 9d

Date:  April 14-15, 1957

Place: Tokyo

Thanks to Jim Z. Yu for translating this commentary into English.]
;B[qd]
C[(* Before WWII, in Japan, the most popular stars in sports were led by Futabayama of sumo \[Japanese wrestling\], Kawakami of baseball, Boriguchi of boxing, and Kimura Meijin of shogi.  While in Go, the most honorable, at the top of the world, were none other than the discoverers of "new opennings", Kitani Minoru and Go Seigen \[Wu\].

(* \[Now it's post-war time.\] Those famous stars \[excluding Wu and Kitani\], at one time or another, have either retired or faded, and only the home run king of baseball, Kawakami, is still able to live up with his fame; he is a rarity.

(* But Wu, since traveling to Japan in 1928 at the age of 14, has held the highest honor in Go, never faded in 30 years. He is truely a genius that would appear only once a few hundred years in Go history. *)]
;W[dc]
C[(* Comparing with competitions in which outcome may be decided in a short burst or a split of second, a Go game takes much longer battles to yield a winner and a loser.  This may have told us why Go players' professional careers are longer than that of other sports. *)]
;B[pq]
C[(* After WWII, Kitani Minoru's records were down for a while, but since he was promoted to 9 dan last year \[1956\], his once unbeatable form appeared again.  Recently he posted an astonishing record of 14 wins and 5 losses.  In major tournaments like the Highest Rank Deciding Matches, Honinbo Matches, and Tokyo Newspaper sponsored Championship Matches, he all recorded triumphs. *)]
;W[oc]
C[(* Is the once overwhelming "Wu-Kitani Era" re-emerging?  Gradually recovered, rising giant Kitani Minoru, challenges Go Seigen for the first time in 13 years.  This balanced struggle is indeed the most remarkable post-war battle since the Wu-Fujisawa 10-game series. *)]
;B[cp]
C[This is the first game between Kitani 9 dan and me \[me = Wu\] after the War.

Before the War, we had a 10-game series when we were 7 dans; later, we also played 4 games.  I remember these 4 games were three games in the Grant Meet (Nihon Kiin's \[Japanese Go Association\] ranked players tournament) and one game sponsored by Yomiuri Newspaper.  Finally in 1944 \[during the War, that is\], we met again in the ranked players tournament when we were 8 dans.  This game now is my first against Mr. Kitani in 13 years.]
;W[po]
C[This game is played in Tokyo Chiyoda District Kioi Street's Fukudake Inn on April 14-15, 1957.  At the same time, the game between Sakata Eio 9 dan and Takagawa Honinbo 8 dan \[another of Strongest Deciding Matches\] is also held here.  That game of theirs will end with Sakata 9 dan, playing Black, winning by 4 points.]
;B[pp]
C[B #7 \[Q4\] is Kitani 9 dan's unique move; other players rarely play it, but Kitani 9 dan has used it a few times in openings like the 1-3-5 \[R16-Q3-C4\] of this game.

B doesn't want to play B\[a\] so that W would play \[Q4\] to form the "avalanche shape", so \[Q4\] move applies the proverb "opponent's key point is also my key point."]
LB[qo:A];
C[If W #8 extends downwards \[R5\] ---

\[Press "shift+down arrow" to switch to this variation, then press "right arrow" to proceed down it.\]

W #8 extending upwards is solid.
]
W[oo]TR[qo]
(;C[B #9 hane is big.]B[qo];
C[W #10 certainly.

\[Switch to the variation to see a reference diagram.\]
]
W[qn]
(;
C[B #11 ---

\[The variation shows the result of playing at the marked point instead of s5.\]
]
B[ro]TR[rp]
(;LB[eq:A][iq:B]
C[W #12 normally plays W\[a\].

Kitani 9 dan said in his post-game review that, if W\[a\], B had prepared to play B\[b\].

In the situation shown in this game, W #12 does have various ways to play.]
W[qg];B[pe]C[B #13 of course.];W[qj]C[W #14 is also normal.];B[ep]
C[B #15 is a big point on the board.

Had #15 been played B\[a\], W would certainly ignore it \[tenuki\] and play W\[b\].]
LB[rn:A][eq:B];W[cf]
C[W #16 is also a big point that cannot miss.  This point and B #15 \[E4\] are miai \[counterbalance moves of same weight\].]
;B[kc]
C[B #17 is an extremely good point.

Next move, W #18, is a difficult one. \[difficult to find a good spot, that is.\]

\[The variation shows what Wu had in mind for a possible W #18.\]
]
(;
C[\[Instead,\] W adopted \[P15\] to attach.  Next move ---

\[Tha main line of play continues with black playing at "A".  Switch to the variation to see what would have happened if black had played at "B".\]
]
W[oe]LB[pg:A][of:B]
(;
C[So B simply presses at \[Q13\] is tesuji.

\[The first variation explains more why B #19 shouldn't play at the marked point.\]

How should W play next move #20? ---

\[The main line continues with W playing at "A".  Variations 2 and 3 show continuaitons where W plays at "B" instead.\]
]
B[pg]LB[pd:A][pf:B]TR[of]
(;C[\[Instead,\] W #20 to squeeze here is a good move.]W[pd];
C[B #21 certainly.]B[qe];
C[\[See variation 1 for what would have happened if W had played hane at the marked point instead of playing at P14.\]
\[The main line of play continues with B playing at "A".  Variation 2 shows what would have happened if B had played at "B" instead.\]

So W #22 extends at \[P14\] is good.
]
W[of]LB[pf:A][qh:B]TR[qc]
(;C[Therefore, B #23 has no choice but \[Q14\].]B[pf];W[qc]
C[W, as a result, gets #24 and ...];B[rc];W[qb]
C[... #26 to settle down the corner.  This local fight thus comes to a pause.]
;B[hc]LB[qh:A][ic:B]
C[If B\[a\] for #27, that would be a good point too, but then W would certainly play W\[b\], and \[b\] is a point that cannot be missed.

B #27 as played \[H17\] is an extremely big point.]
;W[qh]
C[But then W gets #28 \[R12\] key point, and W is in a form to attack B.]
;B[og]
C[If B #29 tenuki \[plays elsewhere\], W\[a\] and B big chunk is captured.]
LB[nh:A];W[rn]
C[W thus finally gets #30 \[S16\].  During this period of conflicts, B never got a chance to play \[S16\].]
;B[np]LB[nq:A][rq:B]
C[B #31 is absolutely the only move.

If B tenuki \[plays elsewhere\], when Wa\[O3\], B has to defend with Bb\[S3\]; letting W to seal B with sente is unbearable.]
;W[iq]
C[W #32 ---

\[The variation shows what would have happened if w had played at the marked point instead of J3.\]
]
TR[oi]
(;B[pi]LB[oi:A][qi:B]
C[B #33 \[Q11\] on one hand prevents W\[a\] that was described previously \[see variation at move #32\], on the other hand aims at B\[b\] cut.]
;W[lq]
C[If W #34 plays at the right side \[around R11\], it would be slow.

\[The variation shows how W would handle it if B plays at the marked point now.\]
]
TR[qi]
(;B[mq]
C[B #35's purpose ---

\[The variation is a reference diagram that shows one of B #35's goals.\]

Also...
]
(;W[lp];B[no]
C[...\[Also\] B now has #37 to get out, aiming at the marked cut at the same time.

At this critical point, W made a mistake in the next move.

W #38 \[W's next move\] ---

\[Here w played at A.  The variations show results of w playing at B -- that would have been correct.\]
]
LB[nn:A][so:B]TR[pn]
(;LB[mn:A]W[nn]
C[In any case, #38 plays in a wrong direction \[summary of variations: here W should \[T5\] first\].

If W\[a\] next, it would *not* be a sente move, thus B wouldn't answer it.

\[The variation shows why W "A" next will not be sente.\]
]
(;B[qi]
C[Because of W's mistake at #38 \[O6\], B finally gets sente, and with that, B cleverly turns to play \[R11\].]
;W[ri]
C[If W can get sente here, W would have W\[a\]-B\[b\]-W\[c\] sequence to extend W's territory at the right side.]
LB[ph:A][oh:B][oi:C];B[rj];W[rk]
C[Had B #41 \[S10\] played \[a\] cut, W would have atari with \[b\].  With this kind of shape, "capture the cut stone" is common sense.]
LB[rh:A][rg:B];
C[It would be awkward for B #43 to atari at \[a\], forcing W\[b\], then B\[c\] to get two W stones. \[Imagine, for example, after B\[c\], W plays \[Q10\].\]]
LB[rh:A][sj:B][rg:C]B[pj];
C[W #44 however ---

\[The variation shows the result of w moving at the marked point.\]
]
W[sj]TR[qk]
(;B[qk];W[rj];
C[To here, W is pressed to a low position.  This is the fault of W #38 \[O6\].

Next ---

\[The variation shows a reference diagram.\]

Therefore, it has become an opening favoring B.
]
B[pk]TR[mn]
(;W[mc]
C[If W #48 tenuki here ---

\[The variation shows the result of a tenuki.\]
]
(;B[fc]
C[B #49 \[F17\] is very good.

On first look \[a\] seems to be the real big point for #49.  But if B\[a\], W will certainly presses at \[F17\].  Then the B teams at upper and right sides would yet have enough eyes, and with two "floating" teams, B would be difficult.

Moreover, now with #49, B creates an opportunity for B\[b\] to attack W.]
LB[ck:A][ce:B];W[ck]
C[Since B gets \[F17\], \[C9\] big point is left for W.];B[gq]
C[And then B #51 \[G3\] becomes clearly the next biggest point.

B #51 could also consider to play at \[N6\] \[see an earlier reference at move #47\], but that would make it complicated, so B #51 "cashes in" with some territory first. \[I guess Wu is saying, since B is leading, no need for him to make it complex.\]]
;LB[mn:A][pm:B][pn:C][cm:D]
C[W\[a\] for #52 is a correct \[text\] move, but if W does play \[a\], it would be no threat to B's corner \[see reference at move #38\].

Also, if W\[a\], B\[b\] to force W\[c\], and B\[d\] would give B a clear lead.

Therefore, even though \[N5\] for W #52 is not solid, W doesn't have many choices.  This is again the fault of #38 \[O6\].

W #52's purpose is to force B repond here, so W can get sente. If B ignores here ---

\[Switch to the variation to see a reference.\]
]
W[mo]
(;C[Therefore, B #53 \[M2\] and ...]B[lr];W[kr];
C[...\[and\] #55 \[N2\] are necessary.

In short, in order to get sente, W has to make a move like #52 \[N5\], although it would leave some weak point.]
B[mr];
C[W #56 is the key point to attack.  No matter what, if W doesn't stick with this B group, W will have no way to win.

#56's purpose is to gain opportunities to protect the marked cut by attacking B team.  Conversely, B can adopt two difference attitudes, either harder or softer, when he chooses his next plays.  In short, this is the first crisis of the game. 

If B chooses the hardest way to play, that is ---

\[Variation 1 shows the result of B playing at "B".\]

Therefore, if B chooses the rough tactic shown above \[in variation 1\], it won't be satisfactory.

I think if B plays at "C" \[variations 2 and 3\] would give W a headache.
]
W[nj]TR[mn]LB[nk:A][pm:B][nl:C]
(;B[nk]C[\[Instead,\] W adopted #57 and ...];W[mk];B[nl]
C[...\[and\] #59.  This is of course the most solid way to play, but ...]
;W[mj]
C[...\[but\] after W #60 connects, W has somewhat achieved his goal of attacking B.]
;B[oj]
C[B #61 ---

\[The variation shown the result of black moving at the marked point.\]
]
TR[on]
(;W[so]
C[W #62 is to gain with sente before sacrificing the marked stones.
]
TR[oo][po];B[rq];W[ml];B[nm];
C[If W #66 connects at \[N6\] ---

\[The variation shown the result of white moving at the marked point instead of N7.\]

(* The big spot at the left side is the key point concerning the trend of both sides' growth.  Here we can see how both sides deliberate intensively over this situation. *)
]
W[mm]TR[mn]
(;B[on];W[mn];LB[qm:A]
C[To here, although W loses two stones, but he gets a spectacular outside thickness in *sente*.  W has managed to gain back what he lost on #38 \[O6\].

After B #69, *if* W can gain with \[a\], seemingly a sente, then turn to left side, that would be even better for W \[than to turn to left directly\].  That is to say ---

\[Variation 1 shows the result of white moving at A; it looks promising.\]
\[Variation 2 shows that black has a way to be safe.\]
]
B[pn]
(;W[cn]
C[\[So Wu said that \[R7\] wouldn't be a sente move for W now.  I thought, "Hey, why bother? \[C6\] is big, get it first anyway." But the fact is, in this level of games, *every* point is important.  If one passes by a sente move, it could be not sente later.\]

Anyway, W finally gets the chance to occupy the big spot at \[C6\], and thus escapes from the previously disadvantageous situation.

At the right side, if B\[a\], W\[b\] cut would take care of it.]
LB[rl:A][ql:B];B[qm]
C[Since W got the big spot at the left edge, letting B to gain here with #71...]
;W[rm];B[ql]C[...and #73 is unavoidable.];W[rl];B[co]
C[B #75 is the way to solidify the corner.  Of course, even if B doesn't play this move, the corner has no problem for now.  But later, when the outside situation changes, then there is a possibility that W will cast in W\[a\].  So B #75 now to prevent it.]
LB[er:A];LB[dn:A][bq:B][bp:C][er:D]
C[White #76 extending downwards decides the growth of both sides' territory.  White extending upwards would be yielding and passive.

Also, White #76 aims at white B, black C, white D.

\[The variation shows what happens if black tenukis and white plays B.\]
]
W[bn]
(;C[B #77 is long-waited severe tesuji.]B[ce];
C[W #78 absolutely *cannot* play at the marked point to give in.
]
W[de]TR[cd];B[cd];
C[W #80 to connect is a strong move.  This move ---

\[The variation shows the result of white playing at the marked point instead of at D14.\]
]
W[df]TR[dd]
(;
C[B #81 ---

\[The variation shows the result of black playing at the marked point instead of D16.\]
]
B[dd]TR[cc]
(;W[ed];
C[Since white #82 blocked, B #83 of course has no choice but to cut, looking forward to a deciding fight.
]
B[ec]TR[ed];C[W #84, too, is the only move.]W[cc];B[bc];W[bb];
C[B #87 ---

\[The variation shows the result of black connecting at the marked point instead of playing at B14.\]
]
B[bf]TR[bd]
(;W[bg]
C[White #88 ---

\[The variation shows the result of white peeping at the marked point instead of playing B13.\]
]
TR[ad]
(;B[bd]
C[B #89 only move.

Next if W\[a\] hane ---

\[The variation shows why white playing hane at "A" would not work.\]
]
LB[ac:A]
(;W[af]
C[W #90 no other choice.  If ---

\[A variation was shown in the comments of move #89.\]

\[The variation shows the result of white playing at the marked point instead of A14.\]
]
TR[ab]
(;B[ab];
C[If W\[a\] connect for #92 ---

\[The variation shows the result of white moving at "A" instead of B19.  This leads to a ko, like the move that white made in the game, but the ko in the game is easier for white.\]
]
LB[cb:A]W[ba]
(;
C[B #93, and the huge ko that would determine the result of this game has started.]
B[ae];W[be];B[ee];W[fd];
C[After B #97, W can't find an appropriate ko threat anywhere on the board, how to save such a dangerous situation?]
B[bf];
C[Under the current circumstances, W #98 is the choice when there's no choice.]
W[ag];B[eb];W[cb];
C[B #101 cuts directly. Good.

This move ---

\[*** 'B' to see a variation.\]]
B[cg]
(;W[gd]
C[W #102.  This point absolutely cannot be passed by.

If W\[a\] for #102, B would follow with \[G16\], and that's a ko threat for B.]
LB[be:A];B[bh]
C[B #103 to pull these few stones \[W \[A13\] etc.\] into the ko is rather unexpected.

This this the second crisis of the game. \[The first was back at move #56.\]

(* Later there will be detailed analysis for B #103.

(* As for this ko, what are the best moves?  Please study it by yourself, then compare your conclusion with B #103. *)]
;W[be];B[rf]
C[B #105 threatens a W team that is worth about 50 points.

(* W to answer a move here seems to be unquestionable, since it's hard to imagine that W would sacrifice such a big group.  But Go Seigen 9 dan's strategies are often unexpected --- *)]
;W[ad]
C[W #106 unhesitatingly finishes the ko.

\[He did!  I couldn't believe it. :)\]]
;B[sh]
C[B #107, and the right W team is dead.  However, I think with the moyo at the middle, W has more than enough to compete with B.

Therefore, B #103 \[B12\] was a questionable move.

In this picture, W G16 \[marked\], extending out, gave W's central moyo a big boost in thickness.  So B's best policy at this point was to shift the emphasis to the middle, damaging W's attempt to control the center. Consequently, B #103 should play as the following.

\[*** 'B' to see reference 1.\]
\[*** 'C' to see reference 2.\]]
MA[gd]
(;LB[ie:A][ic:B][jf:C][ig:D]
C[W #108 wrong order.  Before this, W should play W\[a\], forcing B\[b\], *then* jump at \[a\].

Afterwards if B\[c\] to do a shallow invasion, W\[d\] to fence. This way, from the global point of view, it's a better trade for W.

Also, when W\[a\], if B doesn't play B\[b\] but jumps at \[a\] ---

\[*** 'B' to see a reference.\]]
W[en]
(;
C[Although B #109 is a correct move \[under normal circumstances\], under the current situation, W is better to adopt the following play.

\[*** 'B' to see this variation.\]]
B[hd]
(;W[he];B[ge]
C[If B\[a\] for #111, it would just help W to get \[b\].

B #111 cuts is to apply here some techniques directly, looking for some chances to turn it around.]
LB[ie:A][if:B];W[ie];B[fe];W[ac]
C[Next \[B #115\] ---

\[*** 'B' to see a variation.\]]
(;C[B #115 too deep.

Since B has gone so deep into W's zone...]B[ej];
C[...W #116 has no choice but to surround B and try to kill.]W[gj];
C[B is better to play \[a\] for #117.  Then if W\[b\], B\[c\] to seek a living path in another direction.]
LB[gh:A][fi:B][ig:C]B[el];
C[This B #117 and W #118 exchange \[E8-F7\] is worth a reconsideration \[by B\].  Since inside the influence of W thickness, B #117 could easily lead himself to a trap that he cannot escape. Besides, after W plays #118, when B later tries to invade from the bottom, it will be different. \[It's harder for B to invade from the bottom now, clearly.\]]
W[fm];
C[(* Since it has developed to this situation, the only way for B to win is to desperately find a way to save this team.

(* On the other hand, if W fails to capture B, W would lose.  Therefore W has no choice but to fight his hardest.  It's make-or-break.  As shown, an inspiring and heartbreaking fight has begun. *)]
B[gh];
C[Kitani 9 dan's given time has completely consumed \[each player is given 10 hours; Kitani's clock should show 9 hours 59 minutes, saving the last minute for byo-yomi\], and he doesn't have sufficient time to deliberate.]
W[fi];B[eg];W[fh];B[fg];W[hg];B[gg];W[hh];B[fl]
C[If B #127 plays at \[a\], W would counter with \[b\].

Even if W let B run away towards center, W's right side has a metal-solid wall, and B eventually cannot escape.]
LB[gk:A][gi:B];W[di];B[cj];W[dh]
C[When W #130, if B\[a\] ---

\[*** 'B' to see this variation.\]]
LB[gm:A]
(;B[cl];W[dk];B[dl];W[do];B[dj];MA[dk][do]
C[(* W #132-#134-#136 \[D9-D5-B9\] are related tesuji.  After #134 \[D5\], B's \[lower left\] corner becomes shaky. *)]
W[bk];B[gi];W[fj];B[ek];W[bl];LB[dp:A]
C[(* The middle-game fight has come to a most critical point.  It's such a close and intense melee at the center, B has no chance to defend the weakness at \[a\]. *)]
B[hi];W[dg];B[ji];W[jd]
C[If B ignores W #144, W\[a\] would be unbearable.  No matter how urgent the situation at the center is, B has to answer here a move.]
LB[jc:A];B[jc];W[kh];B[jh]
C[(* Now every move by Kitani 9 dan is under the sound of byo-yomi. *)]
;W[kf];B[hj];W[hl];B[il];W[im]
C[(* Starting from B #115 \[E10, the move that Wu commented as "too deep"\], it has been Kitani 9-dan's unqiue "invading strategy." Often this strategy \[to let opponent form a big frame, then to live a group in this frame\] would terrify the opponent... *)]
;B[hk]
C[(* ...However in this game, W's zone is as solid as a flawless castle, and it's really hard for B to achieve what he has looked for.  Besides, Kitani 9-dan is now short of time, which puts him at a more disadvantageous position. *)]
;W[hm];B[lh];W[ki];B[kg];W[lg];B[jg];
C[When W #160, if B tries B\[a\]-W\[b\]-B\[c\] to make an eye, then W would adopt the techniques shown in the following reference diagram to destory B's \[other\] eye at the bottom \["bottom" here is meant to be around \[K10\]\].

\[*** 'B' to see this reference.\]]
W[le]LB[gf:A][hf:B][ef:C]
(;B[kj];W[li];B[jk];W[hf]
C[W #164 connects to destory the eye at \[a\] is the safest way. Thus, no matter how B struggles, B cannot live now.

(* If W #164 carelessly atari at \[b\] --- *)

\[*** 'B' to see this variation.\]]
LB[ff:A][jl:B]
(;
C[Total of 164 moves.  W wins by resignation.

Time given: 10 hours apiece

Time consumed:

Black: 9 hours 59 minutes

White: 6 hours 26 minutes

\[*** END OF THE GAME ***\]])

(;W[jl]AE[hf]Name[variation]C[(* If W hastily atari at \[K8\]...*)];
B[ih]C[(* B \[J12\] (sente)... *)];W[hf];B[ij]
C[(* ...and B \[J10\] makes a ko. *)

\[*** end of variation.\]
\[*** '\[' to continue.\]])
)

(;
Name[reference]C[\[Suppose after W \[M15\]...\]];B[gf]
C[\[B \[G14\]...\]];W[hf];B[ef]C[\[...and \[E14\] to make an eye...\]]
;W[jl]C[...when W \[K8\] hane...];B[ik]C[...B \[J9\] connects...];
W[jj]
C[...W \[K10\] to deprive an eye from B is a severe tesuji.  With this move, the chances for B to live become slim.

... ...]
;B[kj]C[...then if B \[L10\]...];W[li];B[jk]C[...and \[K9\]...];W[kk]
C[...W \[L10\] is another fatal blow...];B[ij];W[ih]
C[...W \[J12\] denies another eye for B...];B[kl]
C[...even if B \[L8\]...];W[lk];B[jm]
C[...and \[K7\] to capture a W stone...];W[io]
C[...W \[J5\] jumps, and B is still not alive.

\[Blood!\]

\[*** end of reference.\]
\[*** '\[' to continue.\]])
)

(;
Name[variation \(B #131\)]C[\[After W \[D12\]...\]];B[gm]
C[\[If B \[G7\] hane...\]];W[gn];B[em];W[fn]
C[After W \[G6\] and \[F6\]...];B[dn]C[...even if B cuts at \[D6\]...]
;W[dm];B[do];W[dl];B[dj];W[bj]
C[W has \[D7\], \[D8\], and \[B10\] to live at the bottom.  Then...];
B[hn]C[...if B \[H6\] hane...];W[ho]
C[...W \[H5\] also hane, and it works. The result: B is unsuccessful.

\[*** end of variation.\]
\[*** 'g131' to continue.\]])
)

(;
Name[variation \(B #115\)]
C[\[After W \[A17\] capturing B stones...\]];B[hh]
C[If B \[H12\] and...];W[jh];B[fg]
C[...\[and\] \[F13\], this would become a large-scale death-life problem.  To live inside W's surronding stones is probably quite difficult, and in fact, under this situation, even if B doesn't try to struggle to live here, he could instead invade from the lower side to somewhat damage W's territory -- and that could be enough.  But to give up the upper portion is a difficult decision to make, since it's difficult to be sure that B would gain enough from below.

\[*** end of variation\]
\[*** 'g115' to continue.\]])
)

(;B[ie]AE[hd]Name[variation]C[B \[J15\]...];W[jd];B[jc];
C[...W probably would play \[H14\]...]W[hf];B[he];W[gf];
C[...B has \[K13\]...]B[jg];W[ih];B[fn];W[fm];
C[Then \[F6\] and \[G6\] to reduce W territory...]B[gn];W[eo];
C[...to here, B should be quite content with this picture.  Besides, the B territory at upper side is also increased.

Hence, \[J15\] was the key point that neither side should miss.

\[*** end of variation.\]
\[*** '\[' to continue.\]]
B[fp])
)

(;C[\[Say when W \[J15\], B responds with \[E6\]...\]]AB[en]AW[ie]
Name[reference];C[W has \[J18\] tesuji...]W[ib];B[ic];W[jb];B[kb];
W[jc];B[jd];W[hb];B[gc];
C[...to here, it's difficult for this B team to live.

\[*** end of reference.\]
\[*** '\[' to continue.\]]
W[kd])
)

(;
Name[reference 1 \(B #103\)]AE[sh][rf][be][bh][ad]AB[bf][ae]
C[Back to the situation after W #102 \[G16\].];B[hd]LB[bh:A]
C[That \[Wu's opinion on B #103\] is, B doesn't play \[a\], but instead to rise at \[H16\]. ]
;W[be]C[When W \[B15\] captures...];B[sh]
C[...B \[T12\] threats. (If W\[b\] now to finish the ko, B would play \[S14\] to kill W's team.  But this \[situation\] is different from the actual game.  The B\[a\] move in actual game is now at \[H16\] key point, and B is clearly better \[if W continues with the way he played in actual game -- \[A16\] finishes the ko and sacrifices team at right\].) ...]
LB[bh:A][ad:B];W[rf]
C[...so W \[S14\] has to answer the threat of \[T12\]...];B[bf]
C[...B captures...];W[dg]
C[...when W \[D13\]...

\[There is a variation for this move shown in reference 2.\]]
;B[re]C[...B forces another W move here...];W[sg];B[ge]
C[...then from \[G15\] to...];W[fe];B[ff];W[ef];B[gf]
C[...\[to\] \[G14\]...];W[be];B[gc];W[ee];B[bf];W[ch]
C[...after W \[C12\] captures one B stone...];B[dn]
C[...B \[D6\] hane...];W[dm];B[en]
C[...and \[E6\] extends.  Later W needs two moves to clean up the \[upper left\] corner.  This way, W's moyo is greatly shrunken.

\[*** end of reference 1.\]
\[*** '\[' to get back to see reference 2.\]])

(;
Name[reference 2 \(B #103\)]AE[sh][rf][bh][be][ad]AB[ae][bf]
C[\[Again, after W \[G16\]...\]];B[hd]
C[\[The first few moves are the same as reference 1.\]];W[be];B[sh];
W[rf];B[bf];W[ad]
C[Here, if W throws in at \[A16\] \[in reference 1, it's played at \[D13\].\]...]
;B[ac];W[be];B[re];W[sg];B[bf];W[om];B[op];W[be];B[pc];W[ad];B[od];
W[aa];B[nd]
C[...to here, it becomes a big exchange.  B is still better.

\[*** end of reference 2.\]
\[*** '\[' to continue.\]])
)

(;B[gd]AE[cg]Name[variation]C[If W \[G16\] atari instead...];
C[...after W \[F15\]...]W[fe];C[...*then* cuts at \[C13\]...]B[cg];
W[be];B[rf];W[rg];B[bf];W[dg];B[ch];W[be];B[ff];W[ge];B[bf];W[hd];
B[gc];W[be];B[he];W[gf];B[bf];W[hf];B[bh];W[be];
C[When B \[T13\] to threat...]B[sg];
C[...W would decide to finish the ko!...]W[ad];
C[...when B \[T12\] to kill the big W team at the left...]B[sh];
C[...W \[F6\] big jump and with the huge moyo at left, although W sacrificed a big team at left, he has enough to counter B.

\[The key that W *could* sacrifice a huge team at rigth is that W's upper left is now very thick, thus his huge moyo looks very promising.\]

\[*** end of variation.\]
\[*** '\[' to continue.\]]
W[fn])
)

(;W[cb]AE[ba]Name[variation];B[ba];W[ca];
C[...it's a ko.

\[*** end of variation.\]
]B[da])
)

(;W[ab]AE[af]Name[variation]C[If W \[A18\] sagari...];B[cg]
C[...after B cuts here \[C13\]...];W[bh];B[ee];W[fd];B[ef];W[dg];B[ge]
C[...two marked white stones are captured.

\[*** end of variation.\]
]
TR[ed][fd])
)

(;
Name[variation \(W #90\)]C[\[After B \[B16\]...\]];W[ac]
C[If W \[A17\] hane...];B[ab];W[aa];B[ae];W[ch];B[db];W[cb];B[eb];
W[ag];B[af];W[ca]
C[...to here, W makes it seki in *gote*.  W loses big.

\[*** end of variation.\]
])
)

(;W[ad]AE[bg]Name[variation]C[If W \[A16\] peeps...];B[be];W[bg];B[ab]
;W[bd];B[db]C[... ...];W[cb];B[ca];W[ac];B[ba]
C[...to here it's a ko...];W[aa]
C[...when W A19, B has a big ko threat in "A" at the right side.  W thus fails.

\[*** end of variation.\]
]
LB[rf:A])
)

(;B[bd]AE[bf]Name[variation]C[If B connects...];C[W \[A17\] hane...]
W[ac];B[bf];W[ad];B[be];
C[...to here, B is killed.

\[*** end of variation.\]
]W[bg])
)

(;B[cc]AE[dd]Name[variation]
C[If B choose to live the corner with \[C17\] and \[B14\]...];W[ec];
B[bf];W[bg];B[be];
C[...after W \[F16\] hane, W is thick, and it's no good for B.  Also, when time comes, W would have W\[a\]-B\[b\]-W\[c\] to force a ko.

\[*** end of variation.\]
]
LB[cb:A][bb:B][ba:C]W[fd])
)

(;W[dd]AE[df]Name[variation]C[If W connects here...];
C[...B would use \[B14\] to...]B[bf];W[bg];B[cg];W[df];B[bh];W[be];
C[To \[A13\].  After sacrificing the marked stones...
]B[ag]TR[ce][cd]
;W[bd];C[...W's center moyo is gone.

\[*** end of variation.\]
]B[dj])
)

(;C[\[W \[B6\] aims at \[B3\]...\]]Name[reference];
C[\[When W \[B3\]...\]]W[bq];
C[If black blocks at \[C3\] \[instead of \[B4\] shown in the main branch\] ...
]
B[cq];C[...W \[B4\] up to...]W[bp];B[bo];W[ao];B[br];
C[...\[up to\] \[A2\] are extremely big ending moves.

\[*** end of reference.\]
]
W[ar])
)

(;C[\[When B \[Q6\]...\]]Name[reference 1];
C[Suppose W plays \[R7\] to force B to respond...]W[qm];
C[...but B ignores (tenuki) and plays \[C7\], the big point at left side...]
B[cm];C[...say W would then \[P7\]...]W[om];B[op];
C[...and \[Q7\], seemingly to capture the big B team at upper-right.  However ---

\[Back up to the start of the variation and try variation 2 to see how black would save his group.\]
]
W[pm])

(;AB[cm][op][pn][on]AW[qm][om][pm]
C[\[This is the continuation of reference 1 ---

\[Suppose W just played \[Q7\].  The upper-right B team seems to be in great danger...\]]
Name[reference 2]AE[oo][po]CR[pm];C[B has \[S14\]...]B[rf];W[rg];B[rb]
;W[ra];C[...and \[Q12\]...]B[ph];W[nh];B[ng];
C[Although W has \[O12\] and \[M14\] to deprive B's eyeshapes while attacking...]
W[lf];C[...but B has \[T14\], and \[you must see this :-)\]...]B[sf];
W[sb];B[sd]C[\[Here is one eye...\]];W[re];B[sh];W[sg];
C[...\[S8\]!!...]B[rl];W[ql];C[...and \[T8\]!!...]B[sl];W[sm];B[qf];
W[rh];
C[...and \[P8\] makes a second eye! \[first eye at \[S15\]\]

This way, not only B lives here completely, also he already gets the big point \[C7\] at the left.  Therefore, W's hope to gain with sente at \[R7\] \[consult main branch\] is invalid.

\[Here are some explanantions of \[S8\]\[T8\].\]

(* The beauty of \[S8\]\[T8\] is that they prevent W to place a stone at \[Q8\].  If W does play \[Q8\] \[trying to deprive the eye at \[P9\]\], W's whole team is demolished. *)

\[I almost missed it! :-) Indeed, if W \[Q8\] and B \[P9\], then W will be unable to atari \[S8\]\[T8\] stones -- either \[T9\] or \[S7\] would be "suicide." \[S8\]\[T8\] are "standing roosters."\]

\[*** end of reference.\]
]
B[ol])
)

(;W[mn]AE[mm]C[If W \[N6\] connects...];C[...after B \[Q7\] peep...]
B[pm];C[...W \[Q6\] connects...]W[pn];
C[...B gets the big point at the left.  Besides, W leaves a leak at \[N7\].  Thus this W shape is thin.

\[*** end of variation.\]
]
B[cm])
)

(;B[on]AE[oj]Name[variation]
C[A cut at \[P6\] by black would start a large-scale battle...
];W[om]
;B[mn];W[nm];B[ml];W[mm];B[ll];W[ph];B[oh];W[oi];B[li];W[nh];B[ng];
W[mh]
C[...to here, the situation is quite complicated; the outcome is hard to predict for B.

\[So black chose to play \[P10\] instead of a fierce cut at A.\]

\[*** end of variation.\]
]
LB[on:A])
)

(;C[\[Assume B is to adopt the hardest tactic against W \[O10\]...\]]
Name[variation 1 \(B #57\)]CR[nj];C[B \[Q7\] peeks...]B[pm];W[pn];
C[...\[O9\] touches...]B[nk];W[mk];C[...\[N8\] hane...]B[ml];W[nl];
B[nm];W[ok];C[...however, after B \[P8\] atari...]B[ol];
C[...W can struggle with \[N7\] to force a ko...]W[mm];
C[...after B \[O9\] to take the ko...]B[nk];
C[...W has a ko threat at \[P4\] \[assumably more to come\]...]W[op];
C[...so B has to atari first, then...]B[lm];W[mn];
C[...connect to finish the ko...]B[nl];
C[...then W gets \[P3\] up to...]W[oq];B[or];W[nr];B[pr];
C[...\[up to\] \[O3\].  Thus both sides made an exchange.  Since W can next get either \[a\] or \[b\], this result is good for W.

For example, if B plays S3 next to protect the corner, W can play W\[b\]-B\[c\]-W\[d\] "double hane", and B's thickness formed by \[O8\] \[the move that connected the ko\] is gone.

Or, say B ignores the corner and play at the top \[around O13\], what would be the result? ---

\[Continue to see reference\]
\[Conclusion, W is better\]

\[*** end of variation 1.\]
]
LB[rq:A][ng:B][nh:C][mh:D]W[nq];
C[\[Say B ignores the corner and plays around \[O13\]...\]]
Name[reference];C[W tosses in a stone at \[S3\]...]W[rq];B[rr];W[rp];
B[qq];W[qp];B[sq];W[so];
C[...to here, although B is not really killed, but letting W to eat two stones \[previously at R5-S5\] in *sente* is a big loss.

\[*** end of reference.\]
]
B[qs])

(;C[\[After W \[O10\]...\]]Name[variation 2 \(B #57\)]CR[nj];
C[B \[O8\] \[= headache :-)\]...]B[nl];C[...if W \[Q12\] and...]W[ph];
B[oh];C[...\[and\] \[P11\] to cut...]W[oi];C[...B \[O11\] to...]B[ni];
W[oj];B[mi];W[mj];B[lj];W[mk];C[...\[to\] \[N7\] peek and...]B[mm];
W[lk];MA[ph][oi]
C[...\[and\] \[N6\] cut, W would be farily thin.  So W can't hastily cut with \[Q12\] and \[P11\] \[marked stones\].

(After this shape, if W\[a\], B would of course give up the five stones. \[Q9 five stones\])
]
LB[ok:A]B[mn])

(;C[\[Again, after W \[O10\]...\]]Name[variation 3 \(B #57\)]CR[nj];
C[When B \[O8\]...]B[nl];C[...if W \[P8\] to separate...]W[ol];
C[...B starts with \[P7\] up to...]B[om];W[pl];B[nm];W[ql];B[ni];W[mi]
;B[nh];W[mn];B[nk];W[ok];B[mj];W[oj];B[li];W[mk];B[lj];W[lk];
C[...\[up to\] here, B sacrifices four stones \[O8 team\], but...]
B[kk];W[ll];
C[...\[but\] B gets the big point at \[C7\], and it's an easy situation for B.

\[Conclusion: like in variation 2, B comes out better with \[O8\] -- after W \[O10\].\]

\[*** end of variation 3.\]
]
B[cm])
)

(;C[\[Assume B ignores W \[N5\]...\]]Name[reference];
C[W has \[S3\] peek...]W[rq];C[...when B connects...]B[qp];
C[...W \[O2\] another peek.

\[Wu didn't explain further.  The way I see it, W now has either \[N2\] or \[P4\], plus the "bomb" at \[S3\], it *is* a headache to B.\]

\[*** end of reference.\]
]
W[nr])
)

(;
Name[reference]AE[mc]C[\[Suppose W doesn't play N17...\]];B[nd]
C[B immediately peeks at O16...];W[od];B[nc]
C[...and blocks with O17.  These two moves \[O16-O17\] concern the base of both sides; extremely big.

\[*** end of reference.\]
])
)

(;C[\[B can start an attack...\]]Name[reference];C[\[...with N6...\]]
B[mn];W[nm];B[pm];W[pn];
C[To here, although B doesn't quite separate W groups, B is now able to attack this big W team when time comes.

\[*** end of reference.\]
]
B[rl])
)

(;Name[variation]C[If W connects directly at \[R9\]...]W[qk]AE[sj]
AW[qk]AB[rj];C[...B *would* get the two W stones...]B[rh];W[sj];
MA[pj][qk]
C[(* In other words, it's a big difference between with or without B \[Q10\] forcing W \[R9\]. *)

\[So W \[R9\] should \[T10\] capture directly, which W did for his #44.\]

\[*** end of variation.\]
]
B[rg])
)

(;
Name[reference]AW[mn]
C[\[Suppose W has played \[N6\] and B ignores it (tenuki)...\]]CR[mn];
W[rq]C[If W \[S3\] and ...];B[qp];W[qr]C[...\[and\] \[R2\]...];B[pr]
C[...B plays \[Q2\] and there is no play at the corner.

\[Therefore, \[N6\] is not sente.\]

\[*** end of reference.\]
])
)

(;
Name[variation 1 \(W #38\)]C[\[After B \[O5\] ...\]]CR[no];W[so]
C[W hane sente to gain (this play's purpose is to sacrifice \[P5\]\[Q5\] \[marked\] two stones when B cuts at \[Q6\])...]
MA[oo][po];B[rq];W[qk]C[...then extend at \[R9\], and W is settled.];
B[pn]C[If B cuts now...];W[pm];B[on];W[om]
C[...W \[Q7\] and \[P7\] to sacrifice two stones...];B[nn];W[fq]
C[...then turn to \[F3\], an extremely good point...];B[eq];W[er];
B[dr];W[fr]
C[...this way, W gains a good deal.  W is advantageous, needless to say.

In order for B to avoid this...

\[see variation 2 (get back to main branch first)\]

\[*** end of variation 1.\]
])

(;
Name[variation 2 \(W #38\)]C[\[Now look at another variation ...\]]
CR[no];W[so]C[\[W still \[T5\]...\]];B[rq];W[qk]
C[\[...to here is the same as variation 1...\]];B[gq]
C[B, instead of cut \[Q6\], plays \[G3\] here...];W[nn]
C[...then W \[O6\] and...];B[mn];LB[ng:A]W[nm]
C[...\[and\] \[O7\] are appropraite moves.  Now, W\[a\] becomes a severe attack.

\[This is the end of the variation; continuing will show why W "A" is a good attack\]

\[*** end of variation 2.\]
]
;
Name[reference]C[\[Let's see what if W gets chance to play O13...\]]
;W[ng];B[nh];W[mh];B[mg];W[nf];B[ni];W[oj]C[\[*** end of reference\]
])
)

(;
Name[reference]AE[mq]C[\[Imagine B is without \[N3\] ...\]];W[op];
B[oq];W[nq]
C[W can \[P4\] and \[O3\] to cut.

\[This shows the importance of #35 \[N3\] that B played.\]

\[*** end of reference.\]
])
)

(;
Name[variation \(B #35\)]C[\[After W \[M3\]...\]]CR[lq];B[qi]
C[If B pushes and cuts...];W[ri];B[rj];W[rk]
C[W would atari from this side \[S9\], sacrificing the two stones \[R12 and R13\] above.

\[*** end of variation.\]
])
)

(;AW[oi]AE[iq]Name[variation]
C[If W plays \[P11\] to exchange with B \[N13\]...]CR[oi];B[mg];W[iq]
C[*Then* W plays \[J3\]...];B[nd];W[od];B[nf]
C[B would have \[O16\] peep and \[O14\].  Now B's both sides are settled, and B becomes thicker.

\[*** end of variation.\]
])
)

(;AW[qc]AE[of]Name[variation]C[\[If W chooses \[R17\]...\]]CR[qc];
B[of];W[ne];B[rc];W[qb];
C[When B \[R12\] captures one W stone, B gets a good shape.

\[*** end of variation.\]
]
B[qh])

(;C[After W \[P14\]...]Name[variation \(B #23\)]CR[of];
C[If B still wants to capture one W stone with \[R12\]...]B[qh];W[qc];
B[rc];C[...W would have \[R17\] and \[S18\] double hane...]W[rb];B[qb]
;W[pc];B[rd];W[ic];B[ke];
C[...W first gains with \[J17\] and \[J15\]...]W[ie];B[kg];
LB[rg:A][rh:B][ph:C]
C[...then settles the corner with \[T18\].  This way, W is better.

Besides, after this shape, W also has aji of W\[a\]-B\[b\]-W\[c\], aiming to reduce B's liberties from outside, thus sealing B inside.

\[*** end of variation.\]
]
W[sb])
)

(;AB[pg]Name[reference]
C[\[Here is more reason why \[P14\] instead of \[Q13\] for B #19 would be wrong.\]]
;W[qf];LB[of:A][ne:B]
C[Think of this W \[R14\] and B \[Q14\] exchange.  If B played #19 at \[P14\] as shown in the previous variation diagram \[at move #18\], it would be like B now does an exchage of B\[a\] and W\[b\] \[which is clearly bad\].  That's why \[Q14\] for B #19 would be wrong.

\[*** end of variation.\]
]
B[pf]MA[qf])

(;
Name[variation 1 \(W #20\)]C[\[After B \[Q13\]...\]]CR[pg];
C[W \[Q14\] to wedge in seems to be severe, but will it have a satisfactoy result?]
W[pf];B[of];W[qf];C[\[P16\] atari to get out is a mediocre move.]B[od]
;W[ne];B[nd];W[me];B[md];
C[It's certain to get here. This way, not only W enlarged his moyo to the biggest extend, but also W left some aji at \[a\].  If so, W is of course better.  However ---

\[*** variation 3 shows why W \[Q14\] wouldn't work this nicely.\]

\[*** end of variation 1.\]
]
LB[pc:A]W[le])

(;
Name[variation 2 \(W #20\)]C[\[Again, after B \[Q13\]...\]]CR[pg];
C[When W \[Q14\]...]W[pf];
C[B simply \[P16\] to wedge in is tesuji. Then if W\[a\], B\[b\].  This exchange would be in favor of B.

So W \[Q14\] \[for #20\] is not good.

\[*** end of variation.\]
]
LB[of:A][nd:B]B[od])
)

(;
Name[variation \(B #19\)]C[\[After W \[P15\]...\]];
C[If B \[P14\] hane...]B[of];C[W would happily extend at \[O15\]...]
W[ne];B[pg];W[qf];
C[It ends up with this shape; clearly, \[P14\] for B \[#19\] would be a mistake.

\[There is another explanation why \[P14\] would be wrong; see the reference at move #19.\]

\[*** end of variation.\]
]
B[pf])
)

(;
Name[variation \(W #18\)]C[\[Wu says...\]];W[mc]
C[W had considered to play \[N17\] jump...];B[pc]
C[Then B would certainly play \[Q17\]...];W[ic]
C[After that, if W \[J17\] to press...];B[ke];W[me];B[ie];W[gd];B[lg]
C[We can then imagine a picture up to here.  W would look thin, so I didn't adpot this way to play.

\[*** end of variation.\]
\[*** 'g18' to continue.\]])
)

(;AB[rp]AE[ro]Name[variation]C[If B plays \[S4\]...]CR[rp];W[ro];
C[The leak of W to the right is almost covered, so B \[S4\] is not as good as \[S5\], the way actually played.

\[*** end of variation.\]
\[*** '\[' to continue.\]]
B[qp])
)

(;AE[qn][qo][po][oo][pp]Name[reference]
C[\[Japanese editor, Mr. Katsumoto, gives us another angle to look at the way two players have played the corner so far.\]]
;C[\[This is a joseki.\]]W[po];B[qo];W[pn];B[pp];W[qn];
C[(* The way two players have played so far is similar to this joseki. *)

\[*** end of reference.\]
]
B[ro])
)

(;
Name[variation]AW[qo]AE[oo]
C[\[If W extends with \[R5\] instead...\]]CR[qo];B[oo];W[on];B[no];
W[nn];
C[The sequence to here would be certain.  Not only B will have a peek at B\[a\] (followed by W\[b\] connecting and B\[c\] extending), W's \[P6\] and \[O6\] playing on line 6 is also unsatisfactory.

\[*** end of variation.\]
\[Press "shift+left arrow" until you are at the start of the variation, then press "shift+up arrow" to return to the main line of play.\]
]
LB[pm:A][pn:B][qm:C]B[mo])

)