(;FF[4]
CA[UTF-8]
GM[1]
GN[KataGo 7 komi self-play - game 55]
PC[https://online-go.com/review/688006]
PB[KataGo]
PW[KataGo]
BR[9p]
WR[9p]
OT[Error: time control missing]
RE[?]
SZ[19]
KM[7]
RU[NZ]

;B[qd]
;W[dp]
;B[cd]
;W[qp]
;B[fd]
(;W[oc]C[Instead, white says, "I'm immediately going to abandon this stone. The aji left behind by it makes it worthwhile to approach, anyway" (all comments on this game by Anton Christenson, EGF 3d).
]

;B[pe]
;W[nq]
;B[mc]
(;W[pk]C[So it makes sense to me to just leave it at this point and make a more profitable move.
]

;B[cn]
;W[fq]
;B[nd]TR[oc]
(;W[ck]C[So instead white pincers on the left, starting to develop the whole lower side of the board.
]

;B[dh]
(;W[ek]
(;B[cp]C[So black goes for an attach, trying to settle in the corner.
]

;W[cq]
;B[bq]
;W[br]
;B[dq]
;W[cr]
;B[do]
;W[ep]
;B[eo]
;W[fo]
;B[fn]
;W[go]
;B[dj]
;W[em]
;B[en]
;W[ej]
(;B[gn]TR[hm]SQ[cn]SQ[en]SQ[fn]SQ[do]SQ[eo]SQ[cp]SQ[bq]C[Black pushes once here to prevent white from enclosing its group with H7. It's sort of painful to push from behind here on the sixth line. White is making fifth-line area here, but it's worth it for black because now it's a bit harder for white to spend one move here and enclose the black group entirely.
]

;W[ho]
;B[cj]
(;W[eh]
;B[eg]
;W[fh]
;B[dk]
;W[dl]
;B[fl]
(;W[dc]C[Instead, white ignores to play a sort of probing move in this corner. Black almost wants to think about just killing this stone outright, but there's still a bit too much weakness on the outside to do it right now.
]

(;B[fg]TR[dc]SQ[eh]SQ[fh]SQ[ej]SQ[ck]SQ[ek]SQ[dl]SQ[em]C[So black just ignores D17 entirely, saying, "okay, if you want to wreck my corner, you can go ahead, but then I get a very nice attack on the center." Based on the fact that white responds in the game, it seems like this move is sente towards white's center group.
]

(;W[gl]SQ[eh]SQ[fh]XX[gh]SQ[ej]SQ[ck]SQ[ek]SQ[dl]TR[el]SQ[em]C[At first glance, this looks to me like it prevents black from cutting at E8, but it's still perfectly possible for black to cut. Rather, it prevents black from killing everything. Black can maybe kill one part, but now the white group is quite flexible. Black G12 now would not be as severe of an attack as previously.
]

(;B[fb]TR[dc]SQ[oc]C[This is saying that D17 is now completely dead. Black wants to own all of this territory. But just like P17, D17 is not wasted for white. It's going to find some ways to use this stone later as well.
]

(;W[fm]LB[fb:3]LB[fg:1]LB[gl:2]C[Black F18 means that white in effect got two moves in a row down here, so that's kind of nice for white.
]

;B[gh]
;W[cl]
;B[gi]
;W[il]
;B[bo]
(;W[aq]TR[cn]TR[en]TR[fn]TR[gn]TR[bo]TR[do]TR[eo]TR[cp]TR[bq]C[It seems like it's going to be a ko at best for white to kill the black group, and that's maybe not enough at this stage of the game. But it's quite hard for me to understand how it's possible to leave a group like this hanging in this way.
]

(;B[pp]SQ[pk]SQ[il]TR[cn]TR[en]TR[fn]TR[gn]TR[bo]TR[do]TR[eo]SQ[fo]SQ[go]SQ[ho]TR[cp]SQ[dp]SQ[ep]SQ[qp]TR[bq]SQ[fq]SQ[nq]C[To me, it seems like when black plays away, white could find an opportunity to attack the group, but they're both just going to leave it for right now. Black is starting to do something about the huge white moyo.
]

;W[pq]
;B[qo]
(;W[rp]C[Instead, we get exactly the same result with a a sort of funny order of moves.
]

;B[po]
;W[op]
;B[oo]
;W[no]
;B[nn]
(;W[mn]TR[nn]TR[oo]TR[po]TR[qo]TR[pp]C[So white double-hanes instead to keep maximum pressure on this group.
]

;B[mo]
;W[np]
;B[mm]
;W[ln]
(;B[om]TR[on]C[So instead, black plays the tiger's mouth to keep some more eye-shape here. P6 is probably going to be an eye.
]

;W[nl]
;B[nm]
;W[ro]
;B[qq]
(;W[qr]C[KataGo decides to capture it in this way.
]

(;B[ap]
(;W[ar]C[White just connects.
]

;B[lm]
;W[kn]
;B[lk]
;W[km]
;B[bm]
;W[bj]
;B[bi]
(;W[qh]TR[oc]C[So they leave it for now. White extends on the right side, which is quite important just to have visual presence here. This move is also going to be useful later on to activate some aji in the corner.
]

;B[ph]
;W[qi]
;B[jj]TR[dc]TR[oc]SQ[lk]SQ[lm]SQ[mm]SQ[nm]SQ[om]SQ[nn]SQ[oo]SQ[po]SQ[qo]SQ[pp]
(;W[ij]
(;B[hk]TR[eh]TR[fh]TR[bj]TR[ej]TR[ck]TR[ek]TR[cl]TR[dl]TR[gl]TR[em]TR[fm]C[Black plays a really awesome counter-probe here, this double peep. It's such a sharp move. As a human, I would be so concerned about keeping my center points, but KataGo hasn't forgotten about this white group which still has a weakness.
]

(;W[ik]SQ[eh]SQ[fh]SQ[bj]SQ[ej]TR[jj]SQ[ck]SQ[ek]TR[lk]SQ[cl]SQ[dl]SQ[gl]SQ[em]SQ[fm]TR[lm]TR[mm]TR[nm]TR[om]TR[nn]TR[oo]TR[po]TR[qo]TR[pp]C[White has to stay focused on the more important picture, which is the center and the weakness of the black group. But this means that later on, there's a lot of weakness around white's group. Black is going to get something quite big out of it, but we'll see that later.
]

;B[jh]
;W[kc]
;B[lb]
;W[dd]
;B[ce]
;W[hd]
;B[gc]LB[lb:B]TR[dc]LB[kc:A]TR[oc]LB[hd:C]
;W[cc]
;B[bc]
;W[bb]
;B[bd]
;W[ke]
;B[ic]
;W[me]
;B[ng]
;W[lg]
;B[mh]
;W[jg]
;B[ig]
;W[kb]TR[lc]TR[md]
;B[jd]
;W[je]
;B[jf]
;W[li]
;B[kd]
;W[ld]
;B[lc]
;W[if]
;B[kg]
;W[hc]
(;B[jb]C[Black has to come back here.
]

(;W[lh]
;B[pj]
;W[qj]
;B[lj]
;W[mi]
;B[hl]
;W[hj]
;B[oj]
;W[kf]
;B[fp]
;W[gp]
;B[rq]
;W[rr]
;B[fk]SQ[el]TR[fl]SQ[gm]
;W[gm]
;B[jg]
;W[oh]
;B[kl]
;W[ib]
(;B[ka]C[Black has to come back.
]

;W[gb]
;B[ga]
;W[ha]
;B[hm]
;W[gj]
;B[im]
;W[jm]
(;B[gk]
;W[fj]
;B[el]
;W[jl]TR[bj]TR[ck]TR[cl]TR[dl]TR[gl]TR[em]TR[fm]TR[gm]
;B[hb]
;W[hf]
(;B[ia]C[So it seems worthwhile to just remove the ko aji.
]

;W[gg]
;B[gf]
;W[hg]
;B[id]
;W[ok]
;B[nj]
;W[ge]
;B[fe]
;W[mj]
;B[qm]
;W[mk]
;B[ql]
;W[rn]
;B[jk]
;W[ji]
;B[ki]
;W[ii]
(;B[oq]
(;W[sq]TR[oq]C[In the game, white plays here to be really efficient, trying to say, "aha, P3 may be a one-point loss for you possibly, because I was already going to play T3 in some variations!" The only way it can be okay for black to make this exchange is if somehow P3 is going to end up useful, which seems really unlikely at this point, but black is going to make it happen.
]

;B[kj]
;W[rc]
(;B[qc]
(;W[qb]C[So it's really hard for me to know what the logic is here.
]

(;B[hn]TR[io]C[But as usual, at this time black decides to just ignore it and play here instead, which is again really confusing. Why would you play H6 here? This is threatening a hane at J5, but it's sort of confusing. What did black gain, for instance, if white just responds at J6? If this move is actually good for black, shouldn't white have played some atari earlier (for instance after 175. R17)?
]

(;W[kh]TR[hn]C[White doesn't even respond by living in the upper right or anything, white responds by taking this profit, so it's sort of like these are miai for points.
]

(;B[pb]C[Now we continue the sequence in the upper right.
]

;W[ra]
;B[sb]
;W[sc]
;B[pc]
;W[re]
;B[rd]
;W[sd]
;B[rf]
;W[rg]
(;B[qe]C[So atariing here is the best for black.
]

(;W[ab]C[It looks like white is sort of creating ko threats here, preparing for the ko in the upper right, but again this seems a bit strange. Is it really necessary? Aren't these ko threats already big enough? I'm not sure. It can either be that it is necessary to play this way, or that it's fine either way and it doesn't matter, KataGo doesn't care. But we're going to see some exchanges up here.
]

;B[ca]
;W[da]
;B[ba]
;W[dg]
;B[ch]
;W[df]
;B[ef]
;W[de]
;B[bg]
;W[qn]
;B[pn]
;W[sf]
;B[se]
;W[bf]
;B[cf]
;W[re]
;B[io]TR[kh]TR[hn]
;W[ip]
;B[se]
;W[pl]
;B[pm]
;W[re]
;B[jo]
;W[jp]
;B[se]
;W[cg]
;B[af]
;W[re]
;B[ko]
;W[kp]
;B[se]
;W[ml]
;B[ll]
;W[re]
;B[eq]
;W[fp]
;B[se]
;W[ae]
;B[be]
;W[re]
;B[lo]TR[sq]
(;W[lp]
;B[se]
;W[ah]
;B[bh]
;W[re]
;B[qf]
;W[se]
;B[pg]
;W[pi]
;B[nh]
;W[og]
;B[of]
;W[oi]
;B[ne]
;W[mf]
;B[nf]
;W[bp]
;B[ao]
;W[ni]
;B[rk]
;W[rj]
;B[qg]
;W[co]
;B[in]
;W[cp]
;B[ff]
;W[rm]
;B[mp]
;W[mq]
;B[pa]
(;W[si]TR[sj]C[This also prevents a black hane at T10.
]

;B[rh]
;W[sg]
(;B[rl]
(;W[qa]
;B[rb]
;W[sa]
;B[pr]
;W[qq]
;B[ie]
;W[he]
;B[md]
;W[le]
;B[rb]
;W[sb]
;B[ei]
;W[fi]
;B[di]
;W[sm]
;B[sk]
;W[mg]
;B[gd]
;W[hh]
;B[ol]
;W[jn]
;B[dn]
;W[qk]
;B[sj]
;W[sl]C[KataGo ±0

-- chat --
Maharani: Video review by Anton Christenson, EGF 3d: https://www.youtube.com/watch?v=kqGcwLnngGY

]

)(;W[di]
;B[ci]
(;W[ei]
;B[rb]
;W[sa]
;B[rb]
;W[sb]
;B[qa]TR[rb]C[The idea is that if white chooses some other point, black still has the option to throw in like this, which is slightly better.
]

)(;W[qa]
;B[ie]
;W[he]
;B[ei]
;W[fi]
;B[sm]
;W[sn]
;B[md]
;W[le]
;B[sl]
;W[di]
;B[dm]
;W[ei]
;B[gd]
;W[ih]
;B[sk]
;W[sj]C[KataGo gives white -17.5 pp for 266. D11
]

)))(;B[qa]C[I wondered, "why shouldn't black atari here first?"
]

;W[rb]
(;B[rl]C[It would take away one point from white. I think it's definitely possible to do so in sente, but it also takes away another option for black.
]

)(;B[sk]
;W[di]
;B[ci]
;W[ei]
;B[qk]
;W[ol]
;B[md]
;W[le]
;B[ie]
;W[he]
;B[sm]
;W[sn]
;B[gd]
;W[ih]
;B[sl]
;W[jn]
;B[dm]
;W[rl]
;B[sj]
;W[mg]
;B[od]
;W[nk]
;B[ob]C[KataGo gives black -1 pp for 265. R19
]

)))(;W[]
;B[rh]
;W[sg]
;B[sh]
;W[rb]
;B[qa]C[White has to protect against this.
]

))(;W[or]SQ[lo]TR[oq]C[If white had just captured this stone earlier, there would be no way for black to break in here. M5 would have been worth a few points but definitely not a ko threat.
]

)(;W[qf]
;B[mq]C[Obviously, the threat here is that black can attach. This would wreck the whole lower side.
]

(;W[mp]
;B[lp]C[White can't push through because of the liberty shortage here.
]

)(;W[og]
;B[nr]
;W[or]
;B[mp]
;W[oq]
;B[lr]
;W[kr]
;B[ne]C[KataGo gives white -45 pp for 232. R14
]

)))(;W[sf]C[Actually, white doesn't fight the ko right away.
]

;B[se]
;W[qn]
;B[pn]
;W[re]
;B[io]
;W[ip]
;B[se]
;W[ab]
;B[da]
;W[re]
;B[jo]
;W[jp]
;B[se]
;W[ea]
;B[eb]
;W[re]
;B[ko]
;W[kp]
;B[se]
;W[ml]
;B[ll]
;W[re]
;B[lo]
;W[lp]
;B[se]
;W[db]
;B[ca]
;W[re]
;B[qq]
;W[or]
;B[se]
;W[ff]
;B[ef]
;W[re]
;B[eq]
;W[fp]
;B[se]
;W[pl]
;B[pm]
;W[re]
;B[qf]
;W[se]
;B[pg]
;W[bp]
;B[nh]
;W[dm]
;B[bq]
;W[in]C[KataGo gives white -0.4 pp for 190. T14
]

))(;B[sf]C[Here, it would really be temping to try to disconnect this way.
]

;W[qf]
(;B[qe]
;W[sg]
;B[se]
;W[re]C[But then suddenly white is just connected entirely.
]

)(;B[pf]
;W[qg]
;B[io]
;W[jo]
;B[ip]C[KataGo gives black -9 pp for 189. T14
]

)))(;B[io]TR[kh]C[Black is saying, "I'm going to get to reduce some points here eventually." White is saying, "okay, you can have that, but then I'm going to take the points in the center."
]

;W[ip]
;B[pb]
;W[ra]
;B[jo]
;W[jp]
;B[pc]
;W[ko]
;B[sb]
;W[qf]
;B[og]
;W[rd]
;B[re]
;W[qe]
;B[rf]
;W[rg]
;B[pf]
;W[qg]
;B[rb]
;W[sf]
;B[ni]
;W[bp]
;B[ao]
;W[co]
;B[dm]
;W[di]
;B[ci]
;W[ei]
;B[ch]
;W[ne]
;B[oe]C[KataGo gives black -1 pp for 179. J5
]

))(;W[in]TR[hn]C[Did black really gain anything with this exchange? If white plays here, did it change the ko threat situation for the ko that's about to happen in the upper right?
]

(;B[io]C[Is this ko threat slightly bigger than the earlier ko threats available?
]

)(;B[pb]
;W[ra]
;B[sb]
;W[sc]
;B[pc]
;W[re]
;B[rd]
;W[sd]
;B[rf]
;W[rg]
;B[qe]
;W[sf]
;B[se]
;W[pl]
;B[pm]
;W[re]
;B[eq]
;W[fp]
;B[se]
;W[ff]
;B[ef]
;W[re]
;B[io]
;W[jn]
;B[se]
;W[ml]
;B[ll]
;W[re]
;B[fr]
;W[er]
;B[se]
;W[qn]
;B[pn]
;W[re]
;B[hp]
;W[dr]
;B[se]
;W[dg]
;B[ch]
;W[re]
;B[kh]
;W[qf]
;B[ie]
;W[hh]
;B[og]
;W[nh]C[KataGo gives white -1.5 pp for 178. J6
]

)))(;B[pb]TR[oc]C[They're going to play away before finishing this sequence, but let's just look at it. Black sort of needs to hane to prevent white's connection to P17.
]

;W[ra]C[White can threaten to make two eyes.
]

;B[sb]C[Black peeps.
]

;W[sc]
;B[pc]TR[oc]C[Black is going to have to come back and disconnect white from P17.
]

;W[re]C[White is going to be able to jump out here.
]

(;B[rd]
;W[sd]
;B[rf]
;W[rg]
;B[qe]C[We're going to get a ko like this on the right side. That's what's going to happen soon in the game.
]

)(;B[hn]
;W[kh]
;B[rf]
;W[rg]
;B[rd]
;W[sd]
;B[qe]
;W[sf]
;B[se]
;W[dg]
;B[ch]
;W[re]
;B[io]
;W[ip]
;B[se]
;W[ml]
;B[ll]
;W[re]
;B[jo]
;W[jp]
;B[se]
;W[ab]
;B[eb]
;W[re]
;B[ko]
;W[kp]
;B[se]
;W[db]
;B[da]
;W[re]C[KataGo gives black -0.3 pp for 177. Q18
]

)))(;W[in]TR[hn]C[This is surely sente. H6 just looks like such a strange point for black here.
]

(;B[hn]
)(;B[dm]
;W[qb]
;B[pb]
;W[rb]
;B[pc]
;W[re]
;B[rd]
;W[sd]
;B[kh]
;W[qf]
;B[ie]
;W[ff]
;B[ef]C[KataGo gives white -4 pp for 176. J6
]

)))(;B[rd]C[Black can't kill like this.
]

;W[qc]
(;B[pb]TR[oc]TR[qc]TR[rc]C[If black just tries to enclose it completely, white is not alive locally here, but black is to weak on the outside to really make this happen.
]

;W[pc]
;B[ob]
;W[oe]
;B[od]
;W[pd]
;B[qf]
;W[nc]LB[nb:B]LB[md:A]C[Black is falling apart. Maybe A and B next.
]

)(;B[pg]
;W[in]
;B[dm]
;W[nb]
;B[kh]
;W[ie]C[KataGo gives black -8 pp for 175. S16
]

)))(;W[or]C[If white captures here, it's quite obvious that black hasn't gained anything with this exchange. White having a stone here is in no way bad for white. It's not going to be reducing white's liberties in any variation. If my opponent had played P3 as a timesuji, I would respond here without thinking much about it.
]

(;B[rm]
;W[]
;B[sn]TR[qq]TR[rq]C[If we get to here, eventually white is going to have to spend an extra move to capture the two stones.
]

)(;B[kj]
;W[rc]
;B[pb]
;W[in]
;B[dm]
;W[qb]
;B[pc]
;W[re]
;B[rd]
;W[sd]
;B[rf]
;W[rg]
;B[qf]
;W[sf]
;B[kh]
;W[ie]
;B[pg]
;W[pi]
;B[nh]
;W[ni]
;B[og]
;W[oi]
;B[bp]
;W[di]
;B[ci]C[KataGo gives white -4.5 pp for 172. P2
]

)))(;B[kh]
;W[kj]TR[ji]C[White's wedge is saying to black, "you can't save your stones here. I can throw in like this; you have no eyes and you will die."
]

)(;B[kj]TR[jf]TR[ig]TR[jg]TR[kg]TR[gh]TR[jh]TR[gi]SQ[bj]SQ[ck]SQ[cl]SQ[dl]SQ[gl]SQ[em]SQ[fm]SQ[gm]C[Black is going to be forced to connect here and give up these stones. This is part of the profit white got in exchange for dying on the left in gote, making the game still even.
]

;W[rc]
;B[qc]
;W[qb]
;B[pb]
;W[ra]
;B[sb]
;W[sc]
;B[pc]
;W[re]
;B[rf]
;W[qf]
;B[rd]
;W[rg]
;B[sd]
;W[kh]
;B[og]
;W[in]
;B[hn]
;W[ne]
;B[oe]
;W[ml]
;B[ll]
;W[pg]
;B[bp]
;W[di]
;B[ci]
;W[ei]
;B[ch]
;W[nh]
;B[ie]
;W[he]
;B[rk]
;W[rj]
;B[sf]
;W[rm]
;B[mg]
;W[sg]
;B[se]
;W[rl]
;B[qk]
;W[ff]
;B[ef]
;W[sk]C[KataGo gives black -0.4 pp for 171. L10
]

))(;B[]
;W[gb]C[Let's say white captures here.
]

;B[fa]C[Black can't really just back off like this, I think.
]

;W[fe]C[I'm not sure, but this looks kind of annoying to me.
]

))(;B[jl]TR[hn]C[White is saying, "okay, if you try to cut me off right now, then I can cut at H6 and it's not looking so good for you." So black can't do that.
]

;W[hn]
(;B[gk]
;W[fj]
)(;B[og]
;W[pi]
;B[ni]C[KataGo gives black -13 pp for 147. K8
]

)))(;B[hb]C[Black can't capture.
]

;W[id]TR[jd]TR[kd]C[Because then this atari kills two stones.
]

))(;W[ib]TR[kb]TR[kc]C[Later on, white can play this move, which is why it's fine to give up the two stones.
]

;B[ka]
;W[lh]
;B[pl]
;W[gb]
;B[ga]
;W[ha]
;B[hb]
;W[pg]
;B[og]
;W[gb]
;B[hl]
;W[hj]
;B[hb]
;W[id]
;B[jc]
;W[gb]
;B[hf]C[KataGo gives white -7.5 pp for 120. J18
]

))(;B[hb]C[Black is not able to connect here now.
]

;W[id]C[These stones would die.
]

))(;W[hl]C[At this time, it's really not possible for white to connect here.
]

(;B[ik]TR[ij]C[Not only does J10 look kind of stupid now when there's this move available.
]

;W[jk]
;B[jl]
;W[kk]
;B[kl]
;W[kj]
;B[lj]
;W[ki]
;B[jm]
;W[jn]
;B[hj]
;W[im]
;B[ll]
;W[ld]
;B[lc]
;W[kg]
;B[mh]
;W[kc]
;B[fk]
;W[ch]
;B[di]
;W[dg]
;B[df]
;W[cg]C[KataGo gives white -3 pp for 88. H8
]

)(;B[fk]TR[eh]TR[fh]TR[ej]TR[ek]C[But black can also still come out with this stone and get to kill at least the four white stones.
]

)))(;B[ii]C[Let's say black hanes.
]

(;W[kj]
)(;W[ji]C[Then maybe there are a lot more possibilities here for the future. Maybe not right away, maybe after playing some stuff in the center.
]

)(;W[kc]
;B[lb]
;W[dd]
;B[ce]
;W[cc]
;B[eb]
;W[hd]
;B[hb]
;W[le]
;B[jf]
;W[ne]
;B[od]
;W[fe]
;B[jd]
;W[fc]
;B[gc]
;W[gd]
;B[ec]
;W[ed]
;B[jc]
;W[df]
;B[bg]
;W[ch]
;B[dg]
;W[bh]
;B[cg]
;W[di]
;B[pi]
;W[pj]
;B[ni]
;W[mj]
;B[nj]
;W[pg]
;B[og]
;W[oh]
;B[oi]
;W[of]
;B[nh]
;W[pf]
;B[rf]
;W[ie]
;B[qg]
;W[gf]
;B[gg]
;W[je]
;B[kf]
;W[jb]
;B[ib]
;W[ke]
;B[kb]C[KataGo gives black -0.3 pp for 87. J11
]

)))(;W[kc]
;B[jl]
;W[kl]
;B[im]
;W[hl]
;B[jk]
;W[hn]
;B[hm]
;W[gm]
;B[jm]
;W[bp]
;B[rq]
;W[rr]
;B[bq]
;W[ii]
;B[hj]
;W[bp]
;B[mj]
;W[co]
;B[pg]
;W[dd]
;B[cp]
;W[ce]
;B[bq]
;W[dg]
;B[ei]
;W[bp]
;B[be]
;W[co]
;B[cc]
;W[cb]
;B[cp]
;W[hc]C[KataGo gives white -3 pp for 86. L17
]

))(;W[bp]TR[bj]SQ[bm]SQ[cn]SQ[en]SQ[fn]SQ[gn]SQ[bo]SQ[do]SQ[eo]SQ[ap]SQ[cp]C[If white really goes for it, I don't really see how black can live outright here. Maybe I'm missing something, but it's really hard to analyze this with KataGo because it will just say, "ah, black just captures B10 and gives up the group or plays the ko." It doesn't give an outright "black can live" or so, it just doesn't care about it. So maybe that should be my takeaway; I shouldn't be so worried, "is it alive or not?" It's okay if it's a ko.
]

(;B[bk]
;W[bl]
;B[aj]
)(;B[ql]
;W[pl]
;B[pm]
;W[qk]
;B[rl]
;W[qh]
;B[rn]
;W[og]C[KataGo gives white -10 pp for 82. B4
]

)))(;W[]
;B[ar]TR[dr]C[Black winning this ko still has the really big follow-up at D2.
]

)(;W[bp]C[I don't really understand why not to take at least once, make black spend one ko threat?
]

(;B[rq]
;W[rr]
;B[bq]C[It seems like you should always take the ko once. I'm not sure if there's a reason or it just doesn't matter in this case.
]

)(;B[lm]
;W[kn]
;B[bq]
;W[ar]
;B[lk]
;W[km]
;B[bm]
;W[bj]
;B[bi]
;W[cc]
;B[bd]
;W[eb]
;B[bc]
;W[bb]
;B[qm]
;W[ql]
;B[rn]
;W[qh]
;B[qj]
;W[rl]
;B[pj]
;W[oj]
;B[pi]
;W[ol]
;B[ph]
;W[li]
;B[mj]
;W[mi]
;B[ni]C[KataGo gives white -0.1 pp for 74. B4
]

)))(;B[rn]SQ[so]TR[qq]C[If you imagine we have S6 later on, T5 is going to be sente for black.
]

;W[qh]
;B[rl]
;W[rk]
;B[ql]
;W[lm]
;B[ml]
;W[ll]C[KataGo gives black -7.5 pp for 73. S6
]

))(;W[rq]TR[qq]C[White can choose to capture R3 in two different ways. It's very hard for me to really tell here.
]

(;B[qr]
;W[rr]
;B[pr]
;W[oq]TR[mo]C[It's possible that there would be some sort of aji with N5 here, but I have a really hard time seeing it.
]

)(;B[rn]
;W[km]
;B[qr]
;W[rm]
;B[qm]
;W[sn]
;B[ql]
;W[qn]
;B[pr]C[KataGo gives white -9 pp for 72. S3
]

)))(;B[nm]C[Now it's a bit awkward for black. If you connect solidly, you leave behind the Q7 shape point.
]

(;W[pm]C[This can be annoying later on.
]

)(;W[lm]
;B[ll]
;W[kl]
;B[kk]
;W[jl]
;B[mk]
;W[lj]
;B[lk]
;W[ro]
;B[ql]
;W[qk]
;B[rl]
;W[ph]
;B[ap]
;W[ar]
;B[nh]
;W[ni]
;B[mi]C[KataGo gives black -5.5 pp for 67. O7
]

)))(;W[on]TR[nn]TR[oo]TR[po]TR[qo]TR[pp]C[When I see a shape like this, I just sort of assume that this cut is not working, but white could actually cut here. I only realized this after looking through the game several times. I wondered, "wait a minute, why doesn't white cut here?"
]

;B[mo]
;W[np]C[KataGo gives white -22 pp for 62. P6
]

;B[om]
;W[pn]
;B[pl]TR[oo]TR[po]TR[qo]TR[pp]C[One way for black to handle this is to sacrifice the four stones like this.
]

;W[qn]
;B[ok]SQ[fb]SQ[mc]SQ[fd]SQ[nd]SQ[qd]SQ[pe]SQ[eg]SQ[fg]SQ[gh]SQ[gi]TR[pk]XX[oo]XX[po]XX[qo]XX[pp]C[Q9 is completely neutralized and black is building a lot on the upper side, so it can be okay for black to sacrifice.
]

))(;W[op]C[What I would expect here is something like atari and go down.
]

(;B[po]
;W[rp]
)(;B[ap]
;W[bp]
;B[po]
;W[rp]
;B[bq]
;W[ar]
;B[bm]
;W[oo]
;B[ol]
;W[cc]
;B[bd]
;W[bj]
;B[ch]
;W[eb]
;B[ok]
;W[gb]
;B[fc]
;W[fa]C[KataGo gives white -1.5 pp for 56. P4
]

)))(;B[gm]C[This cut is annoying.
]

(;W[gk]C[If white extends here, which is maybe the best shape for the white group, black has an atari.
]

(;B[dm]
;W[el]
;B[cm]C[Suddenly, there's another eye here.
]

)(;B[im]C[There's also some potential to move out to the center area.
]

))(;W[hl]
;B[nr]
;W[or]
;B[oq]
;W[pr]
;B[im]
;W[hm]
;B[ok]
;W[pj]
;B[oj]
;W[pi]C[KataGo gives black -12 pp for 53. G7
]

))(;B[ap]C[Black can play here. White probably can't fight this ko.
]

(;W[ar]C[So let's say white has to come back here, then black got some extra eye shape.
]

)(;W[bp]
;B[pp]
;W[ao]
;B[qq]
;W[qo]
;B[po]
;W[qn]
;B[pn]
;W[qm]
;B[oq]
;W[np]
;B[mn]
;W[rq]C[KataGo gives black -9.5 pp for 53. A4
]

))(;B[bm]TR[bk]C[This move is threatening to connect at B9.
]

;W[bp]
;B[bk]
;W[bl]
;B[al]
;W[bj]
;B[ak]
;W[ch]
;B[di]
;W[bi]
;B[cg]
;W[bg]
;B[cf]
;W[gm]
;B[pp]
;W[pq]C[KataGo gives black -21 pp for 53. B7
]

))(;W[]
;B[ar]TR[bo]SQ[dr]C[If white doesn't respond to B5, now this group is 100 % alive, and black is also threatening to play D2.
]

))(;W[dd]
;B[ce]TR[fb]SQ[dd]XX[fg]C[Normally, you wouldn't be able to play F18 because white could push out like this and get into the center, but now there's nowhere to run thanks to black's F13 stone.
]

(;W[cc]
;B[bc]
;W[bb]C[This is just dead inside, there's not enough space.
]

)(;W[dg]
;B[df]
;W[di]
;B[cg]
;W[ci]
;B[bj]
;W[fm]
;B[hl]
;W[gg]
;B[bi]
;W[hk]
;B[gm]
;W[gk]
;B[gf]
;W[hf]C[KataGo gives white -2 pp for 46. D16
]

)))(;B[el]
(;W[fm]
;B[fk]
;W[fj]C[This is not a ladder.
]

)(;W[cc]
;B[fk]
;W[bd]
;B[pi]
;W[gg]
;B[gf]
;W[hf]
;B[gh]
;W[hg]
;B[ei]
;W[nk]
;B[be]C[KataGo gives black -8 pp for 45. E8
]

)))(;W[cc]
;B[gh]
;W[gj]
;B[ih]
;W[ij]
;B[hk]
;W[hj]
;B[el]
;W[gl]
;B[gk]
;W[fk]
;B[fm]
;W[ik]C[KataGo gives white -2.5 pp for 44. C17
]

))(;B[dd]SQ[dc]TR[cd]TR[fd]C[I think the normal shape instinct in this corner would probably be D16, trying to stay strongly connected with the stones. White can probably make some life, but it's going to be a bit painful and the result is locally quite okay for black. But the situation is a bit different now that black is strong on the outside.
]

;W[dg]
;B[di]
;W[gl]
;B[gk]
;W[fm]
;B[gm]
;W[hl]
;B[cl]
;W[hk]
;B[ec]
;W[el]
;B[dm]
;W[no]
;B[pi]
;W[nj]
;B[qq]
;W[pp]
;B[pq]C[KataGo gives black -1 pp for 43. D16
]

))(;W[el]C[If white responds, this is quite a pleasant exchange for black.
]

;B[bk]
;W[dg]
;B[cl]
;W[cg]
;B[ch]
;W[ee]
;B[ed]
;W[gf]
;B[he]
;W[ig]
;B[jf]
;W[de]
;B[dd]
;W[qh]
;B[jg]
;W[qf]
;B[ih]
;W[ce]C[KataGo gives white -3 pp for 42. E8
]

))(;W[]C[It's still probably not a good idea for black to push and cut. Let's say white just passes.
]

;B[dk]
;W[dl]
(;B[cl]
;W[cm]C[This would still be not very good for black.
]

)(;B[bl]TR[cn]TR[en]TR[fn]TR[gn]TR[do]TR[eo]TR[cp]TR[bq]C[But what black could do is connect underneath like this, so that's at least a sort of safety plan B for this group. There is no way for white to disconnect black here but this would be quite passive. So it's not something black is looking forward to doing in reality but the option is there.
]

(;W[bk]
;B[cl]
)(;W[cl]
;B[bk]
))))(;B[dk]C[There is still this option for black to push and cut here, but it would be a horrible idea to do it right now.
]

;W[dl]
(;B[cl]TR[dk]
;W[cm]
;B[bl]
;W[di]
;B[cj]
;W[ci]TR[di]SQ[dk]SQ[cl]TR[cm]C[White can maybe atari, atari and atari. It's quite horrible for black, so pushing and cutting is not worth thinking about for now.
]

)(;B[cj]
;W[eh]
;B[eg]
;W[fh]
;B[gn]
;W[dg]
;B[ch]
;W[ho]
;B[fl]
;W[gl]
;B[gk]
;W[fm]C[KataGo gives black -10.5 pp for 33. D9
]

)))(;B[cq]C[White doesn't really mind too much black either jumping or sliding into the corner, it's not really anything to worry about in this case.
]

;W[cp]
;B[bp]C[KataGo gives black -4.5 pp for 17. C3
]

)(;B[bp]
;W[cq]
;B[en]
;W[kb]
;B[lb]
;W[dm]
;B[do]
;W[cm]
;B[gn]
;W[hp]
;B[ep]
;W[eq]C[KataGo gives black -3 pp for 17. B4
]

)(;B[en]C[White is saying, "if you want to run out into the center here, be my guest, but then I get to kick now."
]

;W[co]
;B[bn]C[And now it's a bit more awkward because the local shape here is to drop down to the second line. Now black is less out in the center.
]

;W[ei]
;B[ep]
;W[eo]
;B[fo]
;W[do]
;B[fp]
;W[eq]
;B[fm]
;W[gp]
;B[go]
;W[hp]
;B[ho]
;W[ip]
;B[jn]
;W[eh]C[KataGo gives black -2 pp for 17. E6
]

))(;W[co]TR[cn]C[In this sort of situation, I would generally be inclined to kick here to prevent this black group from just jumping into the corner.
]

;B[dn]
(;W[ek]SQ[cn]TR[dn]C[But we could almost think of kicking as helping black to move out faster. D6 is one line closer to the center now compared to C6.
]

)(;W[di]
;B[ch]
;W[fi]
;B[gn]
;W[eh]
;B[ci]
;W[dj]
;B[eg]
;W[fg]
;B[ef]
;W[gp]
;B[in]
;W[gl]C[KataGo gives white -1 pp for 16. C5
]

))(;W[dn]C[This attach would be another option for white. I wanted to ask KataGo about this because I always like to play this attach when I can.
]

(;B[dm]TR[cm]C[As white, I would attach saying, "if black tries to hane here, I have a very pleasant cut." And indeed, it seems like this variation would be a bit good for white.
]

;W[cm]TR[dm]
;B[do]
;W[en]
;B[cp]
;W[eo]
;B[co]
;W[cq]C[White can double-hane here.
]

;B[dq]
;W[ep]
(;B[cr]C[It's not possible for black to atari this way.
]

;W[bq]SQ[cn]SQ[co]SQ[do]SQ[cp]TR[dq]C[D3 or the four stones are about to be captured so black is collapsing here.
]

)(;B[bq]
;W[cr]
;B[bm]
;W[cl]
;B[br]C[So black might have to settle for something like this and it's a bit sad.
]

))(;B[do]TR[dn]TR[dp]C[So instead, black could for instance play this wedge. It's worth keeping in mind that you don't always have to hane this shape.
]

;W[co]C[White can atari.
]

;B[eo]TR[dm]SQ[ep]C[Now black is threatening to ladder at D7 but also to push through at E4.
]

;W[ep]C[So white sort of has to play this slightly awkward shape here to prevent both of those at once.
]

;B[cm]SQ[co]TR[cp]C[Then black can extend, threatening to cut at C4 and capture C5.
]

;W[fn]C[White can squeeze these stones.
]

;B[cp]
;W[bo]
;B[fo]
;W[go]
;B[en]
;W[em]
;B[dm]
;W[fp]
;B[dn]
;W[bp]
;B[fm]
;W[gn]
;B[el]TR[cm]TR[dm]TR[fm]TR[cn]TR[dn]TR[en]TR[do]TR[eo]TR[fo]SQ[ep]SQ[fp]SQ[fq]C[This was one of the variations KataGo gave, and I thought it was really funny that this is still sort of an even-ish result. It's such an ugly shape, for both sides even. Black has this big dumpling but a ponnuki to kind of compensate for it, and white has the F3 empty triangle which looks a bit awkward. This is apparently about even.
]

)(;B[co]C[I think this was also another reasonable option for black.
]

;W[do]
;B[bq]C[Just living in the corner somehow is also quite okay.
]

;W[cq]
;B[cl]
;W[dk]
;B[dl]
;W[el]
;B[ek]
;W[bl]
;B[bm]
;W[bj]C[KataGo gives white -3 pp for 16. D6
]

)))(;W[qc]C[There is some aji there, but it's not really a good idea for white to try to save this stone. It's better to save it for later.
]

;B[rc]
(;W[rb]TR[qc]C[One thing to be looking at throughout the game is this attach and hane. This is not really immediately doing much.
]

;B[pc]
;W[qb]
;B[rd]TR[pc]C[Black can for instance atari and connect here.
]

;W[od]
;B[oe]
(;W[pd]
;B[ob]
;W[pb]
;B[nc]
;W[pc]
;B[nb]C[White is just completely dead - that's no good.
]

)(;W[nb]C[White can play here. Now if black were to respond, things would be a bit different.
]

;B[pb]C[Maybe at this point, black can drop down here.
]

;W[mb]TR[lb]C[Black trying to keep this in would be a bit of an overplay.
]

;B[pd]C[So at this time black can come back.
]

;W[lc]C[White gets to save these stones somehow, but we see that this is not really attractive for white to do right away.
]

;B[jc]C[At most, white is sort of going to get a small group which is just settling locally or even having to fight a bit to survive.
]

))(;W[kb]
;B[lb]
;W[ck]
;B[ch]
;W[ek]
;B[bp]
;W[cq]C[KataGo gives white -5.5 pp for 14. R17
]

)))(;W[oe]TR[oc]C[White could save P17 at this moment and it would be another game.
]

;B[of]
;W[ne]
;B[pc]
;W[pb]
;B[qb]
;W[ld]C[White can run out white these stones but will not make a lot of territory with this group, it will sort of just run away.
]

;B[jc]
;W[lc]
;B[cq]
;W[cp]
;B[dq]
;W[eq]
;B[er]
;W[fr]
;B[fq]
;W[ep]
;B[gr]
;W[dr]
;B[fs]
;W[cr]
;B[di]
;W[ob]
;B[pi]
;W[qk]
;B[je]
;W[hc]
;B[hd]C[KataGo gives white -1.5 pp for 10. P15
]

))(;W[nq]C[White could also make an enclosure here and we could get a very peaceful, almost boring opening (all comments on this game by Anton Christenson, EGF 3d).
]

;B[nd]
(;W[cn]
)(;W[pj]
;B[cn]
;W[fq]
;B[dj]
;W[qg]
;B[en]
;W[lc]
;B[jc]
;W[pd]
;B[qc]C[KataGo gives white -1 pp for 6. O3
]

)))