(;FF[4]
CA[UTF-8]
GM[1]
GN[KataGo 7 komi self-play - game 117]
PC[https://online-go.com/review/1038752]
PB[KataGo]
PW[KataGo]
BR[9p]
WR[9p]
OT[Error: time control missing]
RE[?]
SZ[19]
KM[7]
RU[NZ]

;B[qd]
;W[dd]
;B[dq]
;W[qp]
;B[oq]
;W[po]
;B[cn]
;W[oc]
;B[pe]
;W[lc]TR[oq]C[Probably black will just play in the bottom right now? Personally, I always prefer to make a base for P3, but that's also because I don't really know these variations.

[All comments on this game are a "blind reaction" to the game by GoDave89 (EGF 4d), except bracketed comments by Maharani.\]]
(;B[cf]C[Nope, okay — tenuki again.]
;W[qh]TR[dd]MA[nd]SQ[qd]SQ[pe]TR[oq]C[This is the first slightly unexpected move. I would have thought that counterattacking P3 or responding in the upper right corner would be a little bigger, but putting some pressure on the R16 stones also makes sense. I suspect that black might now just play at O16, or just tenuki.]
(;B[cc]C[For AI, 3-3 is always an option.]
(;W[dc]C[White chooses the other direction.]
;B[cd]
;W[de]SQ[cb]TR[bg]C[Now, typically black plays B13; sometimes C18.]
;B[bg]TR[cb]C[Yeah, this is a really big move. Compared to the variation where black has directly 3-3'd, white doesn't have an immediate threat here with C18. There's not really a special tesuji, so this black shape is nice.]
;W[eg]TR[dc]SQ[lc]TR[dd]MA[ld]MA[nd]TR[de]C[White now cares for the three stones here, and is also threatening to build up the top side a little more; but with the third-line stone at M17, it's not really that easy. Black could now choose to play a move like M16; also just playing O16 might be an option, or just tenukiing again.]
(;B[pc]TR[qd]TR[pe]C[Black chooses to just strengthen the R16 stones now.]
;W[od]
(;B[qf]SQ[dc]SQ[lc]SQ[oc]TR[pc]SQ[dd]SQ[od]TR[qd]SQ[de]TR[pe]SQ[nf]SQ[eg]MA[oq]C[Black chooses to play the vital point of this shape, but gives white some more urgency on the outside, so white might choose to just play O14 next; or she might pincer P3.]
;W[lq]MA[kd]SQ[np]TR[jq]C[Now, black can choose to maybe counter-pincer at K3; maybe move out at O4; or maybe just completely ignore and shoulder-hit at L16.]
(;B[kd]TR[kc]C[Yeah, AI likes to ignore; but this one does make sense. Black wants to ask this question first. Now white can play L17.]
;W[kc]TR[id]C[Yep. And now black might choose to jump at J16.]
;B[id]C[But here, does white just respond?]
(;W[jc]SQ[ja]SQ[ka]SQ[la]SQ[jb]SQ[kb]SQ[lb]TR[id]TR[kd]SQ[eg]MA[ig]C[I would already be happy with this as black, because you did reduce white towards the top side; the L16 stones are more helpful for black than the territory is for white here; and E13 becomes a bit questionable if black just gets reduce this easily. Now black maybe just chooses some random move like the two-space jump at J13?]
(;B[gg]MA[dc]MA[dd]TR[id]MA[de]SQ[hf]MA[eg]C[Oh! Okay, that's interesting; that's not expected. If you play a move like this, you have a really large jump from J16, and that makes it a lot easier for white to split, for example with H14. Black plays this shape, I suppose, to put more pressure on the D16 stones. What could white do if she doesn't split?]
(;W[pk]CR[dc]SQ[gc]CR[dd]MA[gd]CR[de]SQ[he]SQ[hf]CR[eg]TR[gg]SQ[dh]SQ[di]SQ[dj]SQ[dk]SQ[dl]SQ[dm]MA[jq]TR[oq]C[Oh, just tenuki; sure, okay, I guess that makes sense. White is saying that G13 does weaken the D16 group, but she also still has easy ways to split and counterattack, to make a base, and to run to the left side. Q9 is a big move, and in a way, it does put pressure on P3, too. Black might hesitate to move it out now because white has this additional stone, so maybe black just pincers at K3 now? Or maybe he continues on the top side with a jump at G16 to strengthen his shape?]
;B[df]TR[fe]C[Oh, wow, okay. Black pushes and peeps?]
;W[ef]
;B[fe]
(;W[cb]TR[ee]C[So white doesn't just connect; she just settles the group, or at least gets some extra benefit.]
;B[bb]TR[eb]C[Maybe white tiger mouth now?]
;W[eb]TR[ba]SQ[ca]C[White is threatening next to play B19 to make even more eyeshape. Typically, black answers at C19 now.]
;B[ca]TR[eb]SQ[hc]MA[ee]MA[he]C[Now white has gotten E18 in sente, which makes it easier to, for example, connect with a move like H17 later; but I don't think white will play that passively now. Maybe some counterattack around H15 now, or connecting at E15?]
;W[gf]TR[hf]TR[hg]C[Ah okay, yeah; some counterattack. I guess black counters with H14. Maybe extend at H13 in some cases?]
(;B[hf]TR[ge]C[White likely goes down to G15.]
(;W[ge]TR[ff]SQ[fg]C[Now it would be more promising for black to play F14 because white can't really cut at F13.]
;B[ff]TR[gd]C[White will most likely extend at G16.]
(;W[gd]SQ[ge]TR[he]SQ[gf]MA[fg]C[The advantage of extending here compared to turning at H15 is that this group has four liberties now; and black is left with two cuts, anyway. I expect black to connect at F13.]
(;B[fg]SQ[hc]SQ[gd]SQ[jd]TR[ee]SQ[ge]SQ[ie]SQ[gf]TR[hg]C[And now, this is a classic situation of a fighting game. There are immediately dangerous cuts at H13 and E15, and as white, now you have to evaluate who has more problems. Black also has weaknesses at K16 and J15. But white also is not solidly connected at H17, and there are some problems with the liberties of the G14 stones. White might have to connect at E15 now.]
(;W[hg]TR[hf]SQ[if]C[White cuts here. Probably black keeps on fighting on the H14 side? Maybe just extend at J14?]
(;B[if]TR[ig]TR[hh]C[My instinct in these situations as white is always to extend one way or the other.]
(;W[hc]SQ[id]SQ[kd]SQ[ee]SQ[fe]SQ[ff]SQ[hf]SQ[if]SQ[fg]SQ[gg]SQ[hg]MA[jg]TR[hh]C[But since the H12 ladder doesn't work for black and there's not really any satisfying ladder-maker for black on the right side, white comes back here and fixes the weakness. Black still has the cut at E15, but he is also left with the H13 cut on the outside, fixing which is not soluble with one move. But he is probably less worried about the G13 stones and more about the L16 stones, so I would say a diagonal at K13 would make sense for black now.]
(;B[gb]TR[ic]TR[fd]C[Oookaaay..., this is a probing move. Black is threatening to play F16 and something around J17. G18 is actually really annoying for white, but maybe she just connects at J17?]
(;W[ie]SQ[cb]SQ[eb]SQ[dc]SQ[dd]TR[id]TR[kd]SQ[de]MA[fe]MA[ff]MA[hf]MA[if]MA[jf]MA[fg]MA[gg]MA[hh]C[Oh, wow! This also works. White wants black to commit in this area somehow. Which means black might just leave it again, because the L16 stones per se are not that valuable anymore; and it's more about black trying to counterattack the D16 group and handling the outside stones. Black might atari at H12 and then just extend at K14.]
(;B[on]TR[hh]C[Here's the ladder-maker. I expect white to extend at H12.]
;W[ig]SQ[jf]TR[on]MA[oo]C[White pushes. When black made the ladder with P6, either he thought J13 for K14 would be a good exchange; or more likely, he would follow up on P6. I'd suspect that black plays P5 now.]
;B[oo]MA[jf]SQ[pm]LB[qp:B]LB[lq:A]TR[oq]SQ[pq]C[Black says that, "now, I save P3; and A and B become weak, so white better care for them." And if white believes that, she probably wants to play Q3 or jump at Q7. I guess just making sure that the K14 aji is clean is the other option.]
(;W[pq]
;B[nr]TR[pm]MA[on]MA[oo]SQ[lq]MA[oq]C[Black responds. White could just connect at Q7 now, or care for M3. The black group, while now stronger and out, is still without any eyeshape.]
;W[pn]TR[om]SQ[lq]C[White doesn't want to get split here. Extending at P7 would be big for black, but maybe just counter-pincer M3 now.]
(;B[jq]SQ[on]MA[eo]SQ[oo]TR[lq]SQ[oq]SQ[nr]C[Right. Black really doesn't want white to get away with M3, too. K3 is giving an out to black's P3 group, but the main purpose is to put pressure on M3. I suspect white might just treat M3 lightly. Maybe she now plays a move like E5, or follows up on the top side?]
(;W[jf]SQ[gb]TR[id]TR[kd]SQ[fe]SQ[ff]TR[hf]TR[if]SQ[fg]SQ[gg]MA[jo]MA[ko]C[This move is really big. The L16 stones have lost all their value and aji; they can only maybe be used to reduce a bit more. In principle, white now owns the whole top side. And now that white is very strong here, black's H14 stones are gone, and the G13 stones are quite weak (but they're also just worth eight points now if black can't use G18). I would think that black now says, "okay, you didn't follow up on the bottom side; maybe I just play L5 or K5 now."]
(;B[om]TR[on]TR[oo]TR[oq]TR[nr]C[Now black extends, which is not 100 % sente, but strengthens his stones.]
(;W[fm]SQ[fe]SQ[ff]SQ[fg]SQ[gg]MA[dl]TR[cn]MA[fp]TR[dq]TR[jq]C[This move reduces black. I would still say the G13 stones are not that urgent to save. Black could just respond at D8 or F4.]
(;B[jo]MA[cm]MA[dm]SQ[fm]TR[lq]C[Black plays here. Now M3, for all intents and purposes, is pretty much dead. I suspect white will follow up on F7. Maybe at D7 or C7?]
(;W[dp]TR[cp]C[Black probably just blocks.]
;B[cp]TR[eq]C[Then white hanes.]
;W[eq]TR[do]TR[dr]C[Black either descends or ataris.]
;B[do]
;W[ep]TR[dr]C[Now black descends.]
;B[dr]TR[fm]SQ[eo]SQ[ip]SQ[hq]SQ[iq]C[Now F7 is just an additional help for white. I suspect she plays E5 now. Or maybe she makes some base around H3, peeps at J4, or attaches at J3?]
(;W[cq]C[Right, you can cut here first. I always forget about this move.]
;B[cr]
;W[eo]TR[dm]C[Black likely plays D7.]
(;B[dm]TR[cf]TR[df]TR[bg]MA[fk]SQ[fm]TR[cn]TR[en]TR[do]SQ[eo]TR[cp]SQ[dp]SQ[ep]SQ[eq]MA[iq]C[If black can't hane at E6, this other tiger-mouth is the normal shape point. It also raises the left side and gives some additional points, and keeps white's F7 group weaker. I could, again, imagine white attaching at J3 to strengthen that group. Jumping at F9 seems a bit soft.]
(;W[hq]SQ[nf]SQ[qj]TR[om]C[White just makes a base; makes sense. Maybe black now follows up on P7 with a move around R10? Or maybe black just plays O14; but that seal-off is not that big.]
;B[pm]C[Black does do it as directly as I showed earlier, but let's see how AI actually makes use of it.]
;W[qm]C[White blocks.]
;B[ee]TR[pm]TR[qm]C[Q7 was just an exchange... well, that's disappointing. Black finally decides that it's big enough to cut here. There's additional aji looming here, so maybe this is even sente?]
(;W[bq]C[This is sente.]
;B[br]
;W[jr]TR[fc]C[So maybe there's nothing for black around F17? Or maybe white can somehow fix it?]
(;B[pr]
;W[qr]
;B[ql]C[Black activates the aji here. Now we can see how AI does it as just a cut.]
;W[pl]TR[ql]C[I understand this exchange.]
;B[qn]C[But this one?]
;W[rm]TR[qn]C[No idea.]
(;B[kr]TR[pl]TR[ql]TR[rm]TR[qn]MA[hq]SQ[nr]TR[pr]TR[qr]C[KataGo, you're driving me crazy! Why can't you just continue on the right side? What is the benefit of these exchanges? But okay, black returns here. O2 helps; there's actually not that much white can do with H3.]
;W[kq]
;B[jp]
(;W[ir]
;B[lr]
(;W[ml]TR[jf]MA[nf]TR[hg]TR[ig]TR[qh]TR[pk]SQ[pl]SQ[ql]TR[om]TR[pm]MA[ln]TR[on]TR[jo]TR[oo]TR[jp]TR[jq]MA[mq]C[Basically, white just reduces black's outside while growing her own outside, which is relatively strong. There is a slight problem with the two-space jump from Q8, and there's also aji at R8. Maybe black just defends with a move like M6 now? That one seems a bit passive, but white always has N3 in sente and could jump out. Or maybe black counter-reduces at O14.]
(;B[lm]
;W[kn]C[We get to see the AI plan for this move.]
;B[mq]MA[ql]TR[lm]TR[kn]MA[qn]SQ[op]C[Black just fixes like this. It's a reverse-sente move, but then I don't understand the M7 exchange. N3 does fix all of black's issues around P4, but I thought that's what R8 and R6 were already supposed to do.]
;W[er]SQ[ee]SQ[fe]SQ[ff]MA[nf]SQ[fg]SQ[gg]MA[ek]TR[fm]TR[eo]TR[dp]TR[ep]TR[eq]TR[hq]TR[ir]TR[jr]C[White blocks here. This gives her additional endgame benefit, and also makes this group pretty much alive. Now black could just strengthen his G13 stones and build the left side with something like E9, or maybe he is still interested in reducing the right side around O14.]
(;B[gi]TR[ee]TR[fe]SQ[ef]TR[ff]MA[jf]SQ[eg]TR[fg]TR[gg]MA[hg]MA[ig]MA[qh]MA[pk]MA[ml]MA[pl]C[This move is kind of trying to build the left side, strengthening the G13 stones even further without having to capture the E13 stones in a small way. It's also looking at white's center area.]
;W[ek]C[White says, "I don't want you to reduce me AND grow the left side", but this shape is cuttable.]
(;B[rl]C[Black plays this forcing move first.]
;W[rn]
;B[qj]
(;W[oi]TR[qj]TR[ql]TR[rl]C[This makes the life of black's R8 stones more difficult.]
(;B[ph]C[Black attaches on the top side.]
;W[pi]
;B[rh]C[Now black can connect.]
;W[pg]C[But in this way, black is not connected without giving white some additional benefit on the outside.]
;B[qi]
;W[oh]TR[qh]C[White still has to take. Now black could choose to tenuki, but R12 is a really big ko, so I don't suspect it.]
(;B[qg]TR[ij]TR[ck]TR[kk]C[For once, my prediction that AI won't tenuki was correct! White got sente out of this situation. I could imagine white playing some move like L9 or J10 to grow, or maybe go into the left side more around C9.]
(;W[oe]TR[ii]TR[ij]TR[kk]TR[el]TR[ll]TR[km]C[White blocks. Maybe black could jump in at J11 now, or maybe push up at N8? Something like that; or maybe L7, L9, or J10? Or maybe cut at E8 now?]
(;B[gk]C[Ah, okay; black just threatens the cut. I suppose white connects somehow?]
;W[hj]TR[el]C[White peeps first; that kind of connects the E8 cut, but also not really.]
(;B[fc]SQ[ee]SQ[fe]TR[ef]SQ[ff]TR[eg]SQ[fg]SQ[gg]C[This move has a lot of extra benefit aside from just capturing the E13 stones and strengthening the G13 stones.]
(;W[fd]
(;B[gj]
;W[fl]TR[dj]TR[ck]C[Now white protects the cut, and it seems prudent for black to play a move like D10 or C9.]
;B[ed]C[NOW black activates this.]
;W[ec]
;B[fb]C[Is this a ko?]
;W[ln]SQ[da]SQ[ea]MA[im]TR[jo]TR[jp]MA[lp]TR[jq]C[White first threatens to separate the K3 stones. She is trying to create ko threats since she can't win the top left capturing race by just extending at E19, so she'll have to try to play a ko at D19. Black could just capture at M4, or maybe try some weird counterattack around J7.]
(;B[mm]TR[lm]C[Black saves M7. If I was black, I would say, "well, white is trying to create ko threats; I don't want to give white any ko threats"; but if it's a close game, I suppose black doesn't give an inch. Or black is just saying, "I don't fear the ko fight."]
;W[ko]TR[jo]TR[jp]TR[jq]C[I don't think black will just sacrifice the K3 stones, will he?]
;B[in]SQ[om]SQ[pm]SQ[on]SQ[oo]TR[mq]SQ[oq]SQ[kr]SQ[lr]SQ[nr]SQ[pr]C[No. So now it's just a fight, which I don't quite understand. But now N3 does make sense because it makes sure that that black group is pretty much alive.]
;W[km]
;B[kl]
;W[da]TR[gb]C[G18 was always free benefit for black.]
;B[fa]C[Oh, wow; I did not expect this one.]
(;W[jm]C[White just keeps making the fight bigger.]
;B[kp]
;W[hh]C[And then white ignores, anyway... what's going on?]
(;B[ei]C[Black plays a move that doesn't really reduce ko threats, but okay.]
;W[bc]C[And then white captures.]
;B[ea]C[Black ataris?]
(;W[lp]TR[ca]TR[da]SQ[ea]SQ[fa]TR[bb]TR[cb]SQ[fb]SQ[gb]TR[bc]SQ[fc]C[And white returns here to fight the ko. She can always live in the top left corner; black has saved the G18 stones.]
;B[hm]C[Black semi-connects.]
;W[ba]C[White lives.]
;B[ce]
;W[ll]TR[kl]TR[hm]SQ[jm]SQ[km]MA[nm]TR[in]SQ[kn]SQ[ln]TR[jo]SQ[ko]MA[mo]TR[jp]TR[kp]SQ[lp]TR[jq]SQ[kq]SQ[lq]C[Or not, I guess. The top left is apparently always equally as big as the center fight? White cuts off L8 to make the M3 group stronger and also get a potential atari at O7, or maybe N5 as a forcing move? And black's K3 group is not one hundred percent connected.]
;B[hk]TR[jm]TR[km]TR[kn]TR[ln]TR[ko]TR[lp]TR[kq]TR[lq]C[Black connects, and I guess white still has to kind of connect the M3 group.]
;W[ab]C[This move is sente.]
;B[db]TR[ca]SQ[ea]SQ[fa]SQ[fb]SQ[gb]SQ[fc]C[Now the corner is a ko, but black's also saved the G18 stones.]
;W[kk]TR[ca]MA[ed]MA[ee]MA[fe]MA[ff]MA[fg]MA[gg]MA[ei]MA[gi]MA[gj]MA[gk]TR[hk]TR[nl]MA[hm]MA[in]MA[jo]MA[jp]MA[kp]TR[mp]SQ[np]MA[jq]C[Now white has to play here. Now the situation is that black always has the N4 atari, so O4 is a bad move for white in that area; and black is connected up with H9. The ko in the top left is worth something like maybe 18 points. O8 is kind of forcing for black, but I suspect black takes the ko now.]
(;B[nl]
;W[mn]TR[nl]C[Black exchanges once for some reason.]
;B[dk]C[Black tenukis.]
;W[pb]C[White gets this endgame benefit.]
(;B[qs]TR[nl]TR[lm]TR[mm]TR[om]TR[pm]TR[on]TR[oo]TR[mq]TR[oq]TR[kr]TR[lr]TR[nr]TR[pr]C[Black just ignores. This move is really big because it is endgame and also helps this group.]
;W[rs]
;B[ps]TR[qm]TR[rm]TR[pn]TR[rn]TR[po]TR[qp]TR[pq]TR[qr]TR[rs]C[If white ignores now, there's more endgame benefit for black against the corner group.]
;W[rq]TR[qb]C[White just responds, and then black probably responds at R18.]
;B[qb]C[Now white could connect, which is actually really big; or the top left.]
;W[pf]TR[pd]C[This is sente because of the double atari at Q16.]
;B[qe]
;W[dl]
;B[cl]
;W[em]C[This ko is probably heavier for black.]
;B[nk]C[But black doesn't care; he just extends in and lets white reduce a bit more on the left side.]
;W[dj]
;B[ck]
;W[ej]TR[mj]C[Now I suspect maybe black just diagonals in at N10.]
;B[mj]C[This destroys white's middle area quite a bit.]
;W[li]TR[ob]TR[mi]TR[cj]C[White defends. Now black can either push in more at N11, play C10, or atari at P18; those are the big ones. There's also still the ko in the top left; that might be the biggest.]
;B[el]TR[fk]C[Ah, this ko; okay. I don't understand this one. I get that black threatens to cut at F9; maybe to gather ko threats?]
(;W[nm]C[White just captures.]
;B[pj]TR[pk]TR[pl]C[Black captures the Q9 stones in return.]
;W[en]C[White connects.]
;B[mi]C[Black pushes in; that makes sense.]
;W[mh]
(;B[co]C[This move looks really weird, but it also reduces ko threats; so the ko in the top left is still a thing.]
;W[dl]C[White takes back.]
;B[nh]C[Black wedges.]
(;W[lh]TR[ni]C[White just connects. Connecting at O11 now would be sente for black.]
;B[ca]C[Now black starts the ko in the top left.]
;W[dn]C[White gets this atari.]
;B[el]TR[bm]C[Black can respond by taking, which gives white another ko threat at B7 later, potentially.]
;W[sl]
;B[dl]MA[oj]TR[pk]MA[ol]TR[pl]SQ[ql]SQ[rl]C[Black gets to connect the ko now, giving white back the Q9 stones by letting her capture the R8 stones; but black is connected at P10 or P8.]
;W[fk]TR[ql]TR[rl]C[Capturing R8 isn't big enough for white, so black gets to save those, too.]
;B[rk]TR[da]C[Now white probably fights back the D19 ko.]
;W[cj]C[Not at all. They both just ignore the ko in the top left; interesting. Not sure I understand that; it looks quite big.]
;B[ob]
;W[nb]
;B[pa]
;W[da]
;B[hb]C[This is just creating more ko threats here, I suppose.]
(;W[ib]
;B[ca]
;W[rj]C[Now we just start the ko fight full-on.]
;B[sj]
;W[da]
;B[hp]C[I'm not sure this is 100 % sente for black; but if white doesn't answer, black gets a lot of extra benefit, too.]
(;W[ca]TR[cj]TR[dj]TR[ej]TR[ek]TR[fk]TR[fl]TR[em]TR[fm]TR[dn]TR[en]TR[eo]TR[dp]TR[ep]TR[eq]TR[hq]TR[er]TR[ir]TR[jr]C[Can white live with the F7 group? Maybe C19 is white's best chance because she thinks the game is hopeless.]
;B[gq]
(;W[bm]C[White gets this forcing move.]
;B[cm]
;W[gr]
;B[iq]
;W[hr]
;B[gp]C[Black got a lot of forcing moves here.]
;W[es]C[White descends.]
;B[fq]
;W[fr]
(;B[gs]TR[ca]C[Black is claiming white is dead; white is claiming white can live. Or white thought she had to connect at C19 because she didn't have enough ko threats anymore? Black doesn't play G1 if he doesn't think he can kill.]
;W[bk]
;B[bl]
;W[bi]
;B[ch]TR[gn]C[I don't really see how white can live; there's no second eye at G6, anyway, so this just seems like white is dead.]
;W[bp]
;B[bo]
;W[al]
;B[aq]
;W[ci]
;B[aj]C[And that's why white resigns at this point. Sheesh, that did not end very endgame-like. What a wild ride!

[Black 95 % (KataGo)\]

-- chat --
Maharani: Video review by GoDave89 (EGF 4d): https://www.youtube.com/watch?v=XTHN84KO63c
]
)(;B[js]C[I'm surprised it's not this move, but okay; interesting.

[KataGo gives this move -15 pp for black.\]]
;W[fi]C[[KataGo variation.\]]
;B[gn]
;W[fh]
;B[eh]
;W[gh]
;B[hi]
;W[bk]
;B[bl]
;W[bj]
))(;W[gp]C[Can white not cut here?

[KataGo gives this move -14 pp for white.\]]
;B[hr]
;W[iq]
;B[gr]C[Ah, no, white can't even cut.]
(;W[ip]
;B[ho]C[White would be dead; okay, so that is a problem.]
)(;W[ho]C[[KataGo variation.\]]
;B[ip]
;W[gn]
;B[hn]
;W[bk]
)))(;W[gp]TR[id]C[If white responds here, maybe black has more local ko threats? Black also has more ko threats around J16, but white also has some ko threats against the lower left corner.

[KataGo gives this move -2.3 pp for white.\]]
;B[ca]C[[KataGo variation.\]]
;W[si]
;B[ri]
;W[da]
;B[ic]
;W[jd]
;B[ca]
;W[bm]
;B[bl]
;W[da]
;B[ia]
;W[jb]
;B[ca]
;W[bp]
;B[bo]
;W[da]
;B[rr]
;W[sr]
;B[ca]
;W[fi]
;B[fj]
;W[da]
;B[mb]
;W[na]
;B[ca]
;W[al]
;B[ak]
;W[da]
;B[nc]
;W[mc]
;B[ca]
;W[an]
;B[cm]
;W[da]
;B[nd]
;W[ne]
;B[ca]
;W[bj]
;B[aq]
;W[da]
;B[md]
;W[me]
;B[ca]
;W[bk]
;B[bb]
;W[ho]
;B[op]
;W[pp]
;B[hi]
))(;W[na]C[If white now ignores...]
;B[ic]C[... black threatens to play here...]
;W[ib]
;B[jb]TR[ic]TR[id]C[... and here to save the J16 stones.]
))(;W[ni]C[White could cut...

[KataGo gives this move -6.9 pp for white.\]]
(;B[lh]
;W[mg]
;B[lj]
;W[ki]
;B[kj]C[... but that also gives black a lot more threats.]
;W[jj]C[Actually, white can't even block here...]
;B[jk]C[... because of this double atari.]
)(;B[lj]C[[KataGo variation.\]]
;W[ki]
;B[kj]
;W[ji]
;B[ca]
;W[sl]
;B[rk]
;W[da]
;B[ib]
;W[ic]
;B[ca]
;W[rj]
;B[sj]
)))(;B[ib]C[[KataGo gives this move -0.9 pp for black.\]]
(;W[dn]C[White could atari here...]
;B[co]
;W[dl]C[... and here in sente.]
)(;W[bp]C[Or white could still get this atari later.]
)(;W[ic]C[[KataGo variation.\]]
;B[cj]
;W[di]
;B[co]
;W[fk]
;B[ca]
;W[ob]
;B[ad]
;W[bd]
;B[ae]
;W[be]
;B[af]
;W[ch]
;B[bh]
;W[bi]
;B[ci]
)))(;W[hb]C[If white ignores now...]
;B[fk]C[... and black plays here...]
;W[dl]C[... and this becomes a ko, maybe black can win one of the two kos?]
))(;B[qb]C[Basically, if black answers here...

[KataGo gives this move -3.3 pp for black.\]]
(;W[ps]C[... white also has this move, so black has to fight another ko.]
(;B[or]C[If black connects...]
;W[np]C[... maybe there is some additional aji here on the outside?]
;B[op]C[If black has to play here...]
;W[mp]TR[nl]TR[lm]TR[mm]TR[om]TR[pm]TR[on]TR[oo]TR[op]TR[mq]TR[oq]TR[kr]TR[lr]TR[nr]TR[or]TR[pr]C[... and white gets to play here, this whole black group is not alive.]
)(;B[mp]C[Or maybe black ignores and plays here?]
))(;W[pf]C[[KataGo variation.\]]
;B[mp]
;W[rg]
;B[go]
;W[en]
;B[gq]
;W[gp]
;B[hp]
;W[gr]
;B[fp]
;W[fq]
;B[iq]
;W[gp]
;B[nm]
;W[jl]
;B[gq]
;W[hr]
;B[ph]
;W[pd]
;B[qe]
;W[qc]
;B[rc]
;W[nk]
;B[op]
)))(;B[dk]C[Maybe if black plays here now...

[KataGo gives this move -4.6 pp for black.\]]
(;W[dj]C[... white can't play here...]
;B[el]C[... because black is so strong here now, so black can get the whole left side area.]
)(;W[pb]C[[KataGo variation.\]]
;B[qs]
;W[rs]
;B[ps]
;W[rq]
;B[qb]
;W[pd]
;B[qc]
;W[pf]
;B[qe]
;W[ij]
;B[ca]
;W[dl]
;B[cl]
;W[em]
;B[ej]
;W[da]
;B[nl]
;W[nk]
;B[ca]
;W[ck]
;B[dj]
;W[da]
;B[mk]
)))(;W[ba]C[If white now plays here...

[KataGo gives this move -0.5 pp for white.\]]
(;B[ab]C[... I don't think this is working...]
;W[ce]C[... because white can just play this way?]
;B[bd]
;W[be]
;B[ad]
;W[ae]
;B[ac]
;W[hb]C[White can play here...]
;B[af]
;W[gc]C[... and then play here, so white can win it this way.]
)(;B[ce]C[[KataGo variation.\]]
;W[lp]
;B[hm]
;W[ll]
;B[hk]
;W[mn]
;B[ab]
;W[il]
;B[mp]
;W[lo]
;B[gm]
;W[ps]
;B[op]
;W[or]
;B[nm]
;W[nk]
;B[ok]
;W[ol]
;B[nl]
;W[mk]
;B[ks]
;W[cl]
;B[dl]
;W[dk]
;B[ck]
;W[cj]
;B[bk]
;W[di]
;B[bi]
;W[gp]
;B[ii]
;W[jj]
;B[ik]
;W[jk]
;B[ob]
;W[nb]
;B[pb]
)))(;B[ea]C[If black now plays here...]
;W[gh]C[... white can atari...]
;B[fh]
;W[fi]C[... and atari...]
;B[eh]
;W[ei]C[... and atari...]
;B[dh]MA[hh]TR[gi]TR[gj]TR[gk]SQ[in]SQ[jo]SQ[jp]SQ[kp]SQ[lp]SQ[jq]C[... then black really has problems with the separated G11 stones, especially since the K3 stones are not solidly connected yet; so in that regard, I suppose H12 is forcing.]
))(;W[ba]C[If white now captures...

[KataGo gives this move -24 pp for white.\]]
;B[ea]C[... and black plays here...]
(;W[bc]C[... and white ataris here...]
;B[ce]C[... and black ataris...]
;W[ab]
;B[db]C[... then white's group in the corner is alive, okay.]
)(;W[jm]C[[KataGo variation.\]]
;B[hl]
;W[im]
;B[hm]
;W[hn]
;B[kp]
;W[lp]
;B[ho]
)))(;B[lp]C[I suppose if black plays here...

[KataGo gives this move -4.1 pp for black.\]]
(;W[mm]TR[hc]TR[jc]TR[kc]TR[lc]TR[oc]TR[fd]TR[gd]TR[od]TR[ge]TR[ie]TR[oe]TR[gf]TR[jf]TR[hg]TR[ig]TR[pg]TR[oh]TR[oi]TR[pi]TR[hj]TR[pk]TR[ml]TR[pl]TR[kn]TR[ln]C[... and white gets to play some move around here, it makes the center area a lot stronger?]
)(;W[pb]C[[KataGo variation.\]]
;B[mm]
;W[qc]
;B[hb]
;W[ib]
;B[ck]
;W[da]
;B[db]
;W[rg]
)))(;B[ed]C[I suppose, later black can play here...

[KataGo gives this move -0.7 pp for black.\]]
(;W[fb]C[... because white has to play here?]
;B[ec]
;W[gc]
;B[db]C[We shall see; maybe later in the game.]
)(;W[ec]C[[KataGo variation.\]]
;B[gj]
;W[fl]
;B[fb]
;W[pb]
;B[ck]
;W[ln]
;B[hp]
;W[lp]
;B[gq]
;W[gr]
;B[im]
;W[mm]
;B[iq]
;W[hr]
;B[go]
;W[ps]
;B[or]
;W[hl]
;B[gl]
;W[gm]
;B[hm]
;W[il]
;B[jl]
)))(;W[fb]TR[fc]C[F17 for black makes sense...

[KataGo gives this move -26 pp for white.\]]
;B[ec]C[... because this move just works.]
;W[gc]
(;B[db]TR[cb]TR[eb]TR[fb]C[Black could always capture these stones, so there is a lot of aji here.]
)(;B[gj]C[[KataGo variation.\]]
;W[el]
;B[db]
;W[pb]
;B[kk]
;W[jl]
;B[jk]
;W[il]
;B[kl]
;W[jn]
;B[in]
;W[io]
)))(;B[el]C[If black plays this way now...

[KataGo gives this move -28 pp for black.\]]
;W[fk]
;B[fl]
;W[gj]
;B[fj]
;W[hk]
;B[gl]
;W[fi]C[... maybe white can play a variation like this...]
;B[ej]
;W[ei]
;B[dk]
(;W[eh]TR[ee]TR[fe]TR[ff]TR[fg]TR[gg]TR[gi]C[... and then there is this problem for black on the outside with the G13 stones again.]
)(;W[gh]C[[KataGo variation.\]]
;B[eh]
;W[ll]
;B[di]
;W[hi]
;B[ob]
;W[nb]
;B[km]
)))(;B[eh]C[If black now doesn't react...

[KataGo gives this move -18 pp for black.\]]
(;W[pb]C[... white could atari...]
;B[qb]
;W[pd]C[... and atari...]
;B[qc]
;W[pf]C[... and atari...]
;B[qe]TR[hc]TR[jc]TR[kc]TR[lc]TR[oc]TR[gd]TR[id]TR[kd]TR[od]TR[pd]TR[ge]TR[ie]TR[oe]TR[gf]TR[hf]TR[if]TR[jf]TR[pf]TR[hg]TR[ig]TR[pg]TR[oh]TR[qh]TR[oi]TR[pi]C[... and get all of this in endgame later, and it is strengthening the center area; now it's completely white's.]
)(;W[hj]C[[KataGo variation.\]]
;B[fc]
;W[fd]
;B[ed]
;W[ec]
;B[fb]
;W[da]
;B[ba]
;W[pb]
;B[hb]
;W[ib]
;B[ic]
;W[jb]
;B[qb]
))(;B[ll]C[Black's gotta be careful, though. If he pushes up here...

[KataGo gives this move -28 pp for black.\]]
(;W[ln]
;B[mm]
;W[lp]TR[kn]TR[ln]SQ[jo]SQ[jp]SQ[jq]TR[kq]TR[lq]C[... white might be able to save these stones and counterattack the K3 stones, so maybe not this way.]
)(;W[pb]C[[KataGo variation.\]]
;B[qb]
;W[pd]
;B[qc]
;W[pf]
;B[el]
;W[fl]
;B[fk]
;W[fj]
;B[gk]
;W[gj]
;B[ej]
;W[hj]
)))(;W[ii]C[If white plays a move like this...

[KataGo gives this move -12 pp for white.\]]
(;B[oe]C[... black could play here...]
;W[ne]
;B[nf]C[... and here, and this cut...]
;W[of]
(;B[nb]C[... would allow moves like this in endgame...]
;W[ob]
;B[nd]
;W[me]
;B[nc]TR[ob]TR[oc]TR[od]C[... for example, where black could capture some stones.]
)(;B[nd]C[Or black could just straight-up capture like this.]
(;W[me]
)(;W[pb]C[But black has to be careful that white can't play this...]
;B[qb]
;W[me]
(;B[nc]
;W[nb]C[... because in this case, white could connect up.]
)(;B[nb]C[Although this move still works; that's a nice endgame tesuji.]
))))(;B[fc]C[[KataGo variation.\]]
;W[fd]
;B[ed]
;W[ec]
)))(;B[]C[If black passes...]
;W[rg]
;B[qg]
;W[ri]TR[qh]C[... this ko is quite big...]
(;B[ph]C[... and white doesn't stand to lose much on the outside. It's basically a free ko for white for a lot of points.]
)(;B[ck]C[And if white wins the ko...]
;W[sh]C[... not only did white gain more strength and get all these free points...]
;B[dk]
;W[re]
;B[rd]
;W[pb]
;B[qb]
;W[pd]
;B[qc]
;W[ob]TR[qb]TR[pc]TR[qc]TR[qd]TR[rd]C[... but it also threatens this whole corner, which could be a problem for black.]
;B[rf]
;W[sf]
;B[sd]
;W[se]
;B[sb]C[Not quite yet, but it would be a really big reduction at the very least.]
)))(;B[rh]C[If black now attaches here...

[KataGo gives this move -6.2 pp for black.\]]
(;W[rg]
;B[qg]
;W[ri]
;B[ph]
;W[qi]
;B[pi]
;W[pj]
;B[rj]TR[qj]TR[ql]TR[rl]C[... black can't really live with these stones.]
;W[sh]
;B[sj]
;W[si]
;B[rf]
;W[qk]
;B[rk]
;W[sl]C[The capturing race is good for white.]
)(;W[ln]C[[KataGo variation.\]]
;B[mm]
;W[jl]
;B[gk]
;W[em]
;B[dl]
;W[fk]
;B[qi]
;W[lp]
;B[fc]
;W[fd]
;B[ed]
;W[ec]
;B[fb]
;W[da]
;B[db]
;W[ph]
;B[qg]
;W[cb]
;B[ba]
;W[hb]
;B[km]
;W[jn]
;B[jm]
;W[in]
)))(;W[pj]TR[qj]TR[ql]TR[rl]C[White doesn't push here because that makes black's life too easy.

[KataGo gives this move -8 pp for white.\]]
;B[qi]
(;W[pi]
;B[rh]
)(;W[ph]C[[KataGo variation.\]]
;B[rh]
;W[rg]
;B[pi]
;W[oi]
;B[oj]
;W[ni]
;B[og]
;W[of]
;B[ok]
;W[ri]
;B[rj]
;W[oe]
;B[pg]
;W[oh]
;B[ng]
;W[re]
;B[rf]
;W[rc]
;B[rd]
;W[pb]
;B[qc]
;W[qb]
;B[rb]
;W[ra]
;B[sc]
;W[qg]
;B[pf]
;W[mh]
;B[mg]
)))(;B[fl]C[[KataGo gives this move -28 pp for black.\]]
;W[el]
;B[em]
;W[gl]
;B[fk]
;W[fj]
;B[gk]
;W[gj]
;B[hk]
;W[hj]
(;B[ej]
;W[ei]
;B[dj]
;W[fi]TR[ee]TR[fe]TR[ff]TR[fg]TR[gg]TR[gi]C[In this case, white can almost capture the G13 stones. Maybe that's a bit too big when black doesn't really gain much.]
)(;B[fc]C[[KataGo variation.\]]
;W[fd]
;B[eh]
;W[ik]
;B[hl]
;W[il]
;B[hm]
;W[en]
;B[bp]
;W[dj]
;B[ed]
;W[ec]
;B[fb]
;W[da]
;B[db]
;W[ob]
;B[pb]
;W[cb]
;B[ba]
)))(;B[bi]
;W[bp]C[White can later play here...]
;B[co]
(;W[ds]C[... and here, and this is sente.]
;B[bs]
;W[es]
;B[bo]
;W[cs]
;B[ap]TR[bp]TR[bq]TR[cq]C[Now black has to capture these.]
)(;W[bs]C[Maybe there's even this move white can play...]
;B[ar]
;W[ds]C[... to force this way?]
;B[bo]
;W[cs]C[But it's not sente in this case, so maybe not.]
))(;B[eh]TR[ef]TR[eg]C[Earlier, I thought this move would be good to subdue the aji of white's E13 stones.

[KataGo gives this move -12 pp for black.\]]
;W[ln]C[[KataGo variation.\]]
;B[mo]
;W[rg]
;B[rq]
;W[ps]
;B[rn]
;W[rl]
;B[rp]
;W[sn]
;B[ro]
;W[rr]
;B[or]
;W[qq]
;B[sr]
;W[qo]
;B[qk]
;W[qj]
;B[ob]
;W[nb]
))(;B[ok]C[In some cases, black could choose to play something like this.]
;W[ol]
;B[nl]TR[pk]TR[ml]TR[ol]TR[pl]TR[qm]TR[rm]C[White is basically not safely connected here.]
;W[nk]
;B[qk]
;W[oj]
;B[rn]TR[qm]TR[rm]C[And black, for example, captures here now. Although in this case, that's not what white would play.]
)(;B[nk]TR[pk]TR[ml]TR[pl]C[I guess white is not connected either way.]
;W[nl]
;B[ol]
;W[ok]
;B[nj]C[For example, black could cut in sente like this...]
;W[qk]C[... because white now has to return and make sure this aji doesn't work.]
)(;B[ln]C[I wanted to play this way.

[KataGo gives this move -11 pp for black.\]]
;W[er]C[[KataGo variation.\]]
;B[op]
;W[qo]
;B[rl]
;W[rn]
;B[qj]
;W[pi]
;B[pj]
;W[oj]
;B[oi]
;W[nk]
;B[rj]
;W[oh]
))(;W[mq]C[If white now plays here...

[KataGo gives this move -1.2 pp for white.\]]
;B[mr]C[... and black plays here now...]
(;W[op]
;B[np]
;W[nq]C[... in some cases, maybe white could play this kind of ko?]
;B[pp]TR[qn]C[But in this case, since black already has the R6 atari, that's quite a big ko which can't really play. There's almost no ko threat that's big enough; makes sense.]
)(;W[ln]C[[KataGo variation.\]]
;B[op]
;W[qo]
;B[lo]
;W[mo]
;B[ko]
;W[er]
;B[ek]
;W[nl]
;B[mm]
;W[fj]
;B[ej]
;W[mn]
;B[mp]
;W[ei]
;B[eh]
;W[fi]
;B[ol]
;W[ok]
;B[nk]
;W[oi]
;B[ph]
;W[qi]
;B[nj]
;W[kk]
;B[gl]
;W[hk]
;B[gk]
;W[hj]
)))(;W[lr]C[If white now plays here...

[KataGo gives this move -24 pp for white.\]]
;B[ir]C[... black just cuts and counter-ataris...]
(;W[ks]
;B[iq]SQ[qn]MA[op]TR[kq]TR[lq]TR[jr]TR[lr]SQ[pr]SQ[qr]TR[ks]C[... and then maybe even just connects, and white still has to somehow live with this group. Maybe that's also why black exchanged Q2 first, and maybe even why black cut at R6? P4 is kind of sente for black now.]
)(;W[lo]C[[KataGo variation.\]]
;B[ln]
;W[ks]
;B[iq]
;W[mn]
;B[kn]
;W[mm]
;B[mo]
;W[ps]
)))(;B[dk]
;W[kq]TR[om]TR[pm]TR[on]TR[oo]TR[oq]TR[nr]TR[pr]C[Letting white connect back up like this makes the P3 group weak.]
))(;B[fc]C[I do wonder. What if black plays this?

[KataGo gives this move -35 pp for black.\]]
(;W[fb]C[Does white maybe just play here?]
;B[gc]
;W[fd]C[I suppose white can just play it this way.]
;B[ed]
;W[ec]
;B[hb]
;W[ib]
;B[hd]
;W[he]
;B[ic]
;W[jd]C[So there's only a little bit of aji for black.]
)(;W[fd]C[[KataGo variation.\]]
;B[ed]
;W[ec]
;B[fb]
)))(;W[ql]C[I suppose if white connects here...

[KataGo gives this move -6.3 pp for white.\]]
(;B[fc]
;W[fd]
;B[ed]
;W[ec]C[... maybe there is something here for black?]
;B[fb]C[This could be really annoying for white.]
;W[ea]
;B[db]
;W[hb]
;B[da]
;W[ga]
;B[ce]C[Something like this could possibly happen.]
)(;B[iq]C[[KataGo variation.\]]
;W[ek]
;B[eh]
;W[hp]
;B[pr]
;W[qr]
;B[in]
;W[hm]
;B[el]
;W[bq]
;B[br]
;W[fl]
;B[fk]
;W[fj]
;B[gk]
;W[dk]
;B[dl]
)))(;W[eh]C[I could also imagine white just extending here.

[KataGo gives this move -12 pp for white.\]]
;B[pr]C[[KataGo variation.\]]
;W[qr]
;B[fi]
;W[ei]
;B[fj]
;W[ej]
;B[fk]
))(;B[en]C[Now you don't really want to play here as black...

[KataGo gives this move -20 pp for black.\]]
;W[fn]
(;B[em]TR[fm]C[... because in the usual variation, white wouldn't have a stone at F7.]
;W[bq]
;B[br]
;W[el]TR[fm]C[Thanks to F7, white can play this kind of sequence...]
;B[dl]
;W[dk]TR[dl]TR[em]TR[cn]TR[en]TR[do]TR[cp]C[... where the C6 group gets super cramped and black has to be really careful.]
;B[cl]
;W[ck]
;B[bk]C[Because even if black just plays a hane here...]
;W[bl]C[... this one now works for white...]
(;B[bm]
;W[dm]C[... because of the double atari here.]
)(;B[cm]C[And if black has to play here...]
;W[bp]
;B[co]
;W[dn]
;B[dm]TR[cl]TR[dl]TR[cm]TR[em]TR[cn]TR[en]TR[co]TR[do]TR[cp]C[... then later, maybe black really has to be careful with this whole thing. Not quite yet, but the shape is really annoying.]
))(;B[dl]C[[KataGo variation.\]]
;W[em]
))(;B[hq]
;W[dn]C[As black, you also don't really want to give white this kind of forcing move...]
;B[co]
;W[dm]
;B[ho]
;W[bq]
;B[br]
;W[cm]C[... because then, white is again threatening to get this move in sente later.]
;B[bm]
;W[bl]
;B[bn]C[It's really not nice for black to get squeezed like this.]
))(;W[eo]C[[KataGo gives this move -1.1 pp for white.\]]
;B[dm]C[[KataGo variation.\]]
;W[cq]
;B[cr]
;W[hq]
;B[pm]
;W[qm]
;B[pr]
;W[qr]
;B[ee]
;W[ek]
;B[eh]
;W[jr]
;B[ql]
;W[pl]
;B[kr]
;W[kq]
;B[jp]
;W[ir]
;B[lr]
;W[mq]
;B[mr]
;W[ln]
;B[qn]
;W[rm]
;B[op]
;W[qo]
;B[lo]
;W[mo]
;B[ko]
;W[er]
;B[rl]
;W[rn]
;B[qj]
))(;W[lo]TR[lq]C[It's not realistic for white to run out M3 anytime soon...]
;B[kn]
;W[ln]
;B[lm]TR[om]C[... thanks to P7.]
;W[mm]C[... or at least, it would be really, really hard.]
;B[ll]
;W[ml]
;B[mn]C[There's always this cut.]
;W[nn]
;B[mo]
;W[no]
;B[mp]TR[om]TR[on]MA[pn]TR[oo]MA[po]SQ[op]MA[qp]MA[pq]C[There IS some aji with the P6 stones, too; but if black gets to connect at P4, there is maybe more aji for white to worry about on the side.]
)(;W[dm]C[I suspected white might be playing this one now.

[KataGo gives this move -2.2 pp for white.\]]
;B[eo]C[[KataGo variation.\]]
;W[cm]
;B[bn]
;W[ei]
;B[ee]
;W[eh]
;B[gi]
;W[fk]
;B[pr]
;W[qr]
;B[pm]
;W[ps]
;B[hh]
;W[ih]
;B[ii]
;W[hj]
;B[hi]
;W[il]
;B[jk]
;W[jl]
;B[lk]
;W[or]
;B[ql]
;W[qk]
))(;B[]SQ[dm]TR[fm]C[F7 is looking at the D7 shoulder hit.]
;W[dm]
;B[cm]
;W[dk]C[Maybe white creates something in this area.]
;B[eo]C[Maybe white should really claim that if black now plays in this area...]
;W[eh]TR[fe]TR[ff]TR[fg]TR[gg]C[... she can play here and really swallow black's four stones; maybe that's white's angle here.]
))(;W[eh]TR[om]C[But P7 IS quite big.

[KataGo gives this move -15 pp for white.\]]
(;B[pm]
;W[qm]
;B[ql]
;W[pl]
(;B[qk]
;W[qj]
;B[rj]C[For example, something like this could happen.]
;W[qi]
;B[rm]
;W[qn]C[It doesn't quite work for black yet.]
)(;B[rl]C[I guess black can also play this move in sente...]
;W[rm]
;B[qj]C[... and then play here.]
;W[pj]
;B[qi]
;W[pi]
;B[rh]TR[qh]TR[pi]TR[pj]TR[pk]TR[pl]C[Black does have some aji against the R12 stones.]
))(;B[fi]C[[KataGo variation.\]]
;W[dp]
;B[cp]
;W[eq]
;B[do]
;W[ep]
;B[dr]
;W[cq]
;B[cr]
;W[bq]
))(;W[eo]C[I was suggesting this move earlier. I'm not sure why white doesn't play it; this one would also make sense.

[KataGo gives this move -12 pp for white.\]]
;B[dp]C[[KataGo variation.\]]
;W[pr]
;B[pm]
;W[qm]
;B[ql]
))(;B[ko]C[[KataGo gives this move -16 pp for black.\]]
;W[fm]C[[KataGo variation.\]]
;B[eo]
;W[ei]
;B[ee]
;W[eh]
)(;B[jo]C[[KataGo gives this move -4.4 pp for black.\]]
;W[om]C[[KataGo variation.\]]
;B[nm]
;W[ee]
;B[im]
;W[fm]
;B[eo]
;W[fo]
;B[fp]
;W[en]
;B[do]
)(;B[eh]C[[KataGo gives this move -10 pp for black.\]]
;W[om]TR[jf]TR[hg]TR[ig]TR[qh]TR[pk]SQ[on]SQ[oo]SQ[oq]SQ[nr]C[White getting this move would also have been big, kind of creating the center area a bit more and weakening the P3 stones.]
;B[pr]C[[KataGo variation.\]]
;W[qr]
;B[nm]
;W[ln]
;B[lo]
;W[mo]
;B[mp]
;W[jp]
;B[iq]
;W[lp]
))(;W[]TR[nr]C[The benefit of the O2 kosumi is that black now has the option to connect under...]
;B[lr]C[... like this.]
;W[kr]C[It looks like it's working for white to just split here...]
;B[kq]
;W[mr]
;B[mq]
;W[ls]
;B[lp]C[... but not really.]
;W[mp]C[Technically speaking, white can split...]
(;B[lr]C[... but that's still decided by this ko.]
)(;B[nq]
;W[kp]C[Even if white splits like this...]
;B[lo]C[... it's not really clear if white can do much with this.]
))(;W[kq]TR[lr]C[White could play this really ugly bump to make sure that black's connection under at M2 doesn't work...]
;B[jp]TR[cn]TR[dq]TR[jq]C[... but it has the very big disadvantage that now K3 and the left side get stronger.]
)(;W[lo]C[[KataGo gives this move -4.5 pp for white.\]]
(;B[jf]SQ[gb]TR[jd]TR[hf]TR[if]MA[hg]MA[ig]C[At some point, black might decide to pull these stones out and maybe even connect at K16, activating the G18 aji; and then white has to fight with the H13 stones.]
)(;B[fd]C[[KataGo variation.\]]
;W[fc]
;B[jf]
;W[ii]
;B[gc]
;W[hd]
;B[pr]
;W[qr]
;B[gi]
;W[in]
;B[om]
;W[ql]
;B[jo]
;W[lm]
;B[jn]
;W[jm]
)))(;B[om]C[For example, if black now extends, it's not really 100 % sente here...

[KataGo gives this move -0.5 pp for black.\]]
(;W[iq]C[... and white might still be able to play this move.]
)(;W[hp]C[[KataGo variation.\]]
;B[jf]
;W[jg]
;B[pr]
;W[qr]
;B[ql]
;W[pl]
;B[pm]
;W[qm]
;B[kg]
;W[kh]
;B[lg]
;W[lh]
;B[gi]
;W[ii]
;B[mg]
;W[mh]
;B[ng]
;W[ee]
;B[fd]
;W[fc]
)))(;W[jf]TR[on]TR[oo]TR[oq]C[If white now plays here, the lower-right situation is quite annoying for her.

[KataGo gives this move -10 pp for white.\]]
(;B[qm]C[Maybe black can even play a move like this?]
;W[pq]C[And if white now does this...]
;B[or]
;W[pr]C[... and has to live here...]
;B[ro]
;W[rp]TR[qh]TR[pk]C[... something like this, then black can maybe even counterattack the R12 stones at some point.]
)(;B[pn]C[[KataGo variation.\]]
;W[hp]
;B[ee]
;W[fq]
;B[ro]
;W[co]
;B[do]
;W[cp]
;B[dp]
;W[bn]
;B[bo]
;W[bp]
;B[bm]
;W[ao]
)))(;B[ic]C[Now, if black goes down here...

[KataGo gives this move -16 pp for black.\]]
;W[ib]
(;B[hb]C[... and cuts...]
;W[jd]TR[ic]TR[id]C[... white can just capture the two stones.]
)(;B[je]C[[KataGo variation.\]]
;W[hb]
;B[hi]
;W[hh]
;B[gi]
;W[jg]
;B[ji]
;W[jh]
))(;B[fd]C[[KataGo gives this move -4.8 pp for black.\]]
(;W[fc]TR[gc]TR[fd]C[The push-and-cut doesn't quite work for black.]
;B[gc]
;W[hd]
)(;W[gc]C[[KataGo variation.\]]
;B[fc]
;W[fb]
;B[he]
;W[hb]
;B[jh]
;W[hh]
;B[gi]
;W[ii]
;B[kj]
;W[ik]
;B[ji]
;W[jj]
;B[kk]
;W[im]
;B[gl]
)))(;W[]TR[gb]C[G18 is asking how white wants to handle pretty much the following situation.]
;B[ic]
;W[ib]
;B[hb]
;W[jb]
;B[gc]
(;W[hd]
;B[fd]TR[cb]TR[eb]TR[dc]TR[dd]TR[de]C[White's D16 group gets cut off if white is not careful.]
)(;W[fd]
;B[hd]TR[cb]TR[eb]TR[dc]TR[dd]TR[fd]TR[gd]TR[de]TR[ge]TR[gf]C[Of course, this way, white is also cut off.]
)))(;B[eh]TR[fh]C[But black could hane here at any time. White can't cut at F12.

[KataGo gives this move -14 pp for black.\]]
(;W[dg]C[If white goes here...]
;B[ee]SQ[dg]SQ[eh]C[... this is a good exchange for black since white can't extend anymore.]
)(;W[ee]C[If white just connects...]
;B[dh]C[... maybe black can semi-connect on this side.]
)(;W[hh]C[[KataGo variation.\]]
;B[jh]
;W[hj]
;B[jj]
;W[hl]
;B[np]
;W[lo]
))(;B[jg]C[[KataGo gives this move -19 pp for black and only gives it 1 visit out of 436,348 playouts.\]]
))(;W[ig]TR[id]TR[kd]TR[hf]TR[if]C[This kind of helps black in this direction.

[KataGo gives this move -5.7 pp for white.\]]
(;B[gi]C[If black just jumps...]
;W[jf]
;B[hd]C[... I guess this works for white.]
)(;B[jg]C[[KataGo variation.\]]
;W[hi]
;B[ic]
;W[jh]
;B[eh]
;W[jf]
;B[je]
;W[kg]
;B[gb]
;W[hb]
;B[ib]
;W[ee]
;B[fd]
;W[gc]
;B[fi]
;W[fb]
;B[ld]
;W[da]
))(;W[hh]C[Not this way, either. What does white do instead then?

[KataGo gives this move -6 pp for white.\]]
;B[ic]
;W[gb]
;B[jb]
;W[jg]
;B[gi]
;W[hb]
;B[ib]
;W[kf]
;B[ke]
;W[me]
;B[lf]
;W[lg]
;B[nb]
;W[mb]
;B[le]
;W[mf]
;B[md]
;W[ld]
;B[mg]
))(;B[ee]C[Black could counter-cut now.

[KataGo gives this move -3 pp for black.\]]
(;W[ig]SQ[id]SQ[kd]TR[ef]SQ[hf]TR[eg]C[I suppose white can also give up the E13 stones and then have a good fight on the outside, so the benefit for black on the left side is not big enough.]
)(;W[hh]C[[KataGo variation.\]]
;B[if]
;W[hc]
;B[gi]
;W[jh]
;B[ij]
;W[fm]
;B[hm]
;W[gk]
;B[hi]
;W[jg]
;B[ki]
;W[dm]
;B[eo]
;W[cm]
;B[bn]
;W[ej]
;B[oo]
;W[on]
;B[no]
;W[iq]
;B[kp]
;W[lp]
;B[nn]
;W[ko]
)))(;W[hc]C[Basically, white doesn't want to fix here immediately...

[KataGo gives this move -21 pp for white.\]]
(;B[ig]TR[id]TR[kd]SQ[ee]TR[fe]TR[ff]TR[hf]TR[fg]TR[gg]C[... because if black gets to play here, now his outside is quite connected and strong, and white is still left with the E15 weakness.]
)(;B[ee]C[[KataGo variation.\]]
;W[fm]
;B[np]
;W[pr]
;B[on]
;W[lo]
;B[mn]
)))(;B[hg]TR[fg]C[I don't suppose black will allow the F13 cut...]
;W[fg]SQ[cb]SQ[eb]SQ[dc]SQ[dd]SQ[gd]SQ[de]TR[fe]SQ[ge]SQ[ef]TR[ff]SQ[gf]SQ[eg]C[... because the F15 stones are now just completely lost; there's almost no aji left, and that makes this white group alive.]
))(;W[fg]C[Now white can't cut here...]
;B[gd]TR[ge]TR[gf]C[... because otherwise, the G14 stones are just captured in a ladder.]
)(;W[he]TR[ge]TR[gf]C[White could save the stones like this, but that's also really horrible shape. If white plays here, this group only has three liberties.

[KataGo gives this move -25 pp for white.\]]
(;B[ie]
;W[hg]C[If white ataris...]
;B[gd]TR[hf]C[... black might just give up the H14 stone.]
;W[if]
;B[hd]
;W[hf]
;B[fg]
;W[fd]C[White could cut...]
(;B[hh]TR[ge]TR[he]TR[gf]TR[hf]TR[if]TR[hg]C[... but this is really horrible shape.]
)(;B[ee]C[Also, there's still this cut.]
))(;B[fg]C[[KataGo variation.\]]
;W[gc]
;B[ee]
;W[fm]
;B[np]
;W[pr]
;B[on]
;W[lo]
;B[mn]
;W[pm]
;B[ln]
;W[jo]
;B[hq]
;W[kn]
)))(;W[ff]TR[ef]TR[eg]C[If white now plays here, she has a kind of empty triangle, which is not really that efficient a shape.

[KataGo gives this move -6.1 pp for white.\]]
;B[ge]C[[KataGo variation.\]]
;W[hg]
;B[if]
;W[gh]
;B[jh]
;W[hc]
;B[hh]
;W[fg]
;B[hi]
;W[fj]
;B[np]
;W[pr]
;B[iq]
;W[lo]
;B[nn]
;W[lm]
;B[om]
;W[pl]
))(;B[ff]C[If black now cuts, that would be exactly what white wants...

[KataGo gives this move -2.8 pp for black.\]]
;W[fg]
;B[hf]
;W[gh]C[... because white does get quite a few stones outside in sente...]
(;B[hg]C[... and black is not really all too happy with this.]
;W[hh]C[White probably extends here. She has to read...]
;B[fh]C[... but this cut is not really a threat...]
(;W[ee]TR[fe]TR[ff]C[... because if white ever connects here, it also threatens to capture the F15 stones, and then black has more problems.]
)(;W[fi]TR[fh]C[Also, the ladder is just working, so the F12 cut doesn't work for black right now.]
))(;B[ge]C[[KataGo variation.\]]
;W[hh]
;B[ee]
;W[hb]
;B[fc]
;W[fb]
;B[gc]
;W[mo]
;B[gb]
;W[ga]
;B[fk]
;W[fq]
;B[hk]
;W[jh]
;B[db]
;W[ec]
;B[hq]
;W[go]
;B[io]
)))(;W[ee]SQ[dc]SQ[dd]SQ[de]TR[fe]SQ[ef]SQ[eg]C[The reason white doesn't just connect here is because in general, F15 is more outside, so E15's not really helpful; and white's shape is also kind of clunky, so she would prefer not to play a move that doesn't have any point value.

[KataGo gives this move -1.5 pp for white.\]]
;B[iq]C[[KataGo variation.\]]
;W[mo]
;B[io]
;W[fm]
;B[ej]
;W[dp]
;B[cp]
;W[eq]
;B[do]
;W[ep]
;B[dr]
;W[eo]
;B[dm]
;W[im]
;B[ko]
))(;W[hf]C[[KataGo gives this move -4.7 pp for white.\]]
(;B[gf]
;W[he]TR[id]TR[gf]TR[gg]C[This would just be a very crude way, but black can't really stay connected.]
)(;B[hg]C[[KataGo variation.\]]
;W[gf]
;B[if]
;W[ig]
;B[jf]
;W[fg]
;B[gi]
;W[kg]
;B[jg]
;W[ih]
;B[hi]
;W[jh]
;B[kh]
;W[ki]
;B[lh]
;W[jd]
;B[li]
;W[ie]
;B[kf]
;W[jj]
;B[kj]
))(;W[np]C[I would have thought maybe if you tenuki, you'd play something like this?

[KataGo gives this move -0.7 pp for white.\]]
;B[pk]C[[KataGo variation.\]]
;W[fq]
;B[df]
;W[ef]
;B[fe]
;W[cb]
;B[bb]
;W[eb]
;B[ca]
;W[hc]
;B[qq]
;W[rq]
;B[rr]
;W[qr]
;B[pq]
;W[sr]
;B[oo]
;W[op]
;B[eo]
;W[dr]
)(;W[lo]C[Or something like this?

[KataGo gives this move -1.7 pp for white.\]]
;B[pk]C[[KataGo variation.\]]
;W[fq]
;B[df]
;W[ef]
;B[fe]
;W[ee]
;B[gi]
;W[jd]
;B[qr]
;W[rr]
;B[hq]
;W[go]
;B[io]
;W[dp]
;B[cp]
;W[cq]
;B[cr]
;W[bq]
;B[do]
;W[ep]
)(;W[gd]
;B[qj]SQ[qh]TR[pk]C[In the game, white Q9 prevents black from pincering and attacking R12 while also destroying something on the right side with R10.]
))(;B[ig]TR[id]C[The reason I wanted to play a move like this is because it has a better relation to J16. In general, if you have a two-space jump or a large knight's move, it's harder to separate this cleanly.

[KataGo gives this move -2 pp for black.\]]
(;W[if]C[Even though there are possible cuts for white here...]
;B[jf]
;W[je]
;B[ie]
;W[hf]C[... it's not that easy for white either...]
;B[jd]TR[jf]C[... because having the extra stone at K14 might help black.]
;W[ke]
;B[ld]TR[jc]TR[kc]TR[lc]C[Also, black could put pressure on the M17 stones, or just escape here. I'm not sure how, exactly; maybe there's a better way for black to handle this?]
)(;W[pk]C[[KataGo variation.\]]
;B[iq]
;W[lo]
;B[qr]
;W[rr]
;B[qq]
;W[rq]
;B[oo]
;W[on]
)))(;W[kf]C[The problem is sometimes that white might counterattack here.

[KataGo gives this move -6.1 pp for white.\]]
;B[jc]C[[KataGo variation.\]]
;W[ld]
;B[if]
;W[np]
))(;B[np]C[If black now moves this out, maybe this fight also helps white to create something on the outside.

[KataGo gives this move -0.4 pp for black.\]]
(;W[nf]C[I don't know if this is the best way for white.]
;B[kd]C[But now, if black shoulder-hits...]
;W[kf]C[... white might already be strong enough to counterattack immediately.]
)(;W[pq]C[[KataGo variation.\]]
;B[kd]
;W[ld]
;B[on]
;W[lo]
;B[ln]
;W[ko]
;B[mn]
;W[pl]
;B[kf]
;W[kn]
;B[ll]
;W[kl]
)))(;B[oe]C[Black could play here.

[KataGo gives this move -2.7 pp for black.\]]
(;W[qf]TR[pc]TR[qd]TR[oe]TR[pe]SQ[qe]C[This angle point is a bit annoying for the black shape. Black wouldn't usually answer at R15 now.]
;B[lq]
;W[pd]C[But something like this could happen...]
;B[qe]
;W[qc]
;B[rc]
;W[pb]C[... where white now takes away the base.]
(;B[qb]C[But this could still happen.]
)(;B[rb]TR[rc]TR[qd]TR[oe]TR[pe]TR[qe]C[Or like this; just, this black group is not settled yet.]
))(;W[oh]C[[KataGo variation.\]]
;B[lq]
;W[le]
;B[gg]
;W[qf]
)))(;B[ld]SQ[dc]SQ[dd]TR[qd]SQ[de]TR[pe]SQ[eg]C[Black would kind of try to help the R16 stones and also reduce the top.

[KataGo gives this move -4.9 pp for black.\]]
;W[kc]C[[KataGo variation.\]]
;B[md]
;W[mc]
;B[lq]
;W[qf]
;B[je]
)(;B[lq]C[It's just that at some point, if black tenukis too much, for example here...

[KataGo gives this move -1.9 pp for black.\]]
(;W[ne]TR[dc]TR[lc]TR[oc]TR[dd]SQ[qd]TR[de]SQ[pe]TR[eg]C[... white might even choose to play a move like this and build the top while attacking the R16 stones.]
)(;W[qb]C[[KataGo variation.\]]
;B[nd]
;W[nc]
;B[md]
;W[ld]
;B[mf]
;W[pk]
;B[kf]
;W[je]
;B[om]
;W[np]
;B[qm]
)))(;W[cd]C[White can block here.

[KataGo gives this move -0.6 pp for white.\]]
;B[dc]
;W[ed]SQ[cd]SQ[dd]TR[cf]C[And then it comes down to some variation where C14 is more or less almost an attack on the D16 stones.]
;B[bd]
;W[be]
;B[bc]
(;W[bf]
;B[cg]
;W[ci]
;B[bg]
;W[ce]
;B[fc]TR[cd]TR[dd]TR[ed]TR[be]TR[ce]TR[bf]C[For example, something like this is something I remember; not sure if it's still the most modern version of the joseki. But here, black basically makes white's stones heavy first and takes the corner, which is a bit better than if black had to pincer first.]
)(;W[np]C[[KataGo variation.\]]
;B[nq]
;W[dp]
;B[cp]
;W[mp]
;B[lq]
;W[eq]
;B[dr]
;W[ep]
;B[do]
;W[eo]
;B[en]
;W[fn]
;B[em]
;W[fm]
;B[el]
;W[cq]
;B[er]
;W[cr]
;B[gr]
;W[bp]
;B[co]
)))(;B[fc]TR[dd]C[It used to be that if white ignored her D16 stone, black would double-approach for sure.

[KataGo gives this move -1.8 percentage points winrate (pp) for black.\]]
(;W[fd]
;B[gd]
;W[fe]
;B[ec]
;W[dc]C[And then we would have some variation like this.]
)(;W[df]C[[KataGo variation.\]]
;B[dg]
;W[ef]
;B[ce]
;W[cd]
;B[eg]
;W[ic]
;B[hd]
;W[ec]
;B[id]
;W[fd]
;B[jc]
;W[qb]
;B[md]
;W[mc]
;B[nd]
;W[kc]
;B[pi]
;W[qi]
)))(;B[cc]C[[KataGo gives this move -1.3 pp for black.\]]
;W[dc]
;B[cd]
;W[de]
;B[bf]
;W[qh]
(;B[cg]TR[cb]TR[be]TR[dg]C[This shape is weaker compared to the game variation because white still has some attachments available.]
)(;B[pc]C[[KataGo variation.\]]
;W[od]
;B[qf]
;W[lq]
;B[kd]
;W[kc]
;B[id]
;W[ib]
;B[np]
;W[lo]
;B[qr]
;W[no]
;B[rq]
;W[fq]
;B[eo]
)))