Toroidal go - catching up in the endgame

As a follow up to the previous post, I would like to take an in-depth look into another game I played with Sighris, in 2017.

In this game Sighris played a violent cut early on (black 13) which I handled quite badly, choosing the wrong direction several times. Sighris made the most of this opportunity and played well; the position was good for him after black 33. After white 44 the endgame started and I was behind. I began to catch up little by little, thanks to some endgame mistakes by Sighris. In particular black 83 and 87 were mistakes (exercise: find better moves for Black). After 88 I ended up winning this closely-fought game by 3.5 points.

(;GM[1]FF[4]CA[UTF-8]AP[go-variants-transformer]ST[2]SZ[11]KM[4.5]HA[0]DT[2017-09-27~2018-01-19]PB[Sighris]PW[Malcolm Schonfield]EV[go19.in.TOROID.85]SO[http://littlegolem.net/jsp/game/game.jsp?gid=1892972];B[ff];W[ib];B[eb];W[jg];LB[bg:A][bc:B]C[This seems a little slow to me. My first ideas for candidate moves here are ‘A’ and ‘B’.]B[fi];W[cf];B[ba];W[bj];B[ik];W[bc];B[jd];W[kb](;C[I hadn’t expected this rather violent cut. It ended up giving Black a good result. But I don’t think it’s a good move - see the variation for white 14.]B[jb](;C[I think this move was a mistake.]W[jc];LB[ic:A][kc:B]TR[jd]C[The marked stone seems to be well placed. There is bad aji for White at ‘A’ and ‘B’ that can’t be mended with a single move.]B[ja](;C[This was also a mistake I think.]W[ic];C[It seems a bit too hard to handle for White here.]B[kc];W[kd];C[It feels tough for White now.]B[ac](;C[Maybe this is an overplay.]W[ab];B[ad];W[ke];B[ae];W[af];LB[cd:A]C[I think the next few moves by Black are good. With Black my first instinct here would be to play at ‘A’ instead, but maybe that isn’t as good.]B[kf];W[je];B[bf];W[ag];B[be];W[bg];B[ce];W[bb];C[This seems to be very good for Black. There is too much bad aji in White’s position.]B[de](;W[ca];B[ak](;LB[jd:A][kf:B]C[I probably played this as an attempt to mitigate the bad aji in ‘A’ and ‘B’. But it feels a bit unpromising.]W[hf];LB[ie:A]C[This position feels good for Black to me. There is bad aji around ‘A’, and the four stones are captured. The rest of the game was a difficult struggle for me. I was behind. I tried to surround some territory and tried to make enough points elsewhere, but it felt pretty hopeless.]B[cb];W[ck];B[aj];W[db];B[cc];W[dc];B[cd];C[I think the endgame phase starts here. Black is ahead.]W[ec];B[cg];W[fa];B[ek];W[ea];B[kh];W[kg];B[ch];W[ai](;C[I think this cut didn’t help Black much.]B[bi];W[ci];B[bh];W[ah];B[di];LB[cj:A]C[This move means Black can’t capture at ‘A’ straight away - see the variation.]W[kj](;B[kk];LB[cj:A]C[Again, this move means Black ‘A’ doesn’t work.]W[jj](;B[bk];W[cj](;C[This was necessary - see the variation.]B[jk];W[ij];C[Threatening a snapback - see the variation for white 66.]B[dj](;W[da];B[hk];W[gg];B[fg];W[ee];B[ef];W[gd];B[gh];W[hh];B[hi];W[ii];B[hj](;LB[ih:A]C[I should probably have played ‘A’ here instead. But maybe I felt that playing simply like that wouldn’t be enough.]W[fk];B[ih];W[jh];B[hg];W[gf](;C[Black missed a chance here. I think I remember, I was very relieved when I saw this move as I had been worried about the atari (see the variation). Still, even with this missed opportunity, I think Black may well be slightly ahead here.]B[fj];W[ig];B[hh];W[fe](;C[I think this was the losing move. ]B[ga];W[gb];B[hb];W[hc];B[ha];W[dk];B[ej];W[dd];MA[jd][eb][cf][bb][kh]C[move 95 White stones captured by Black: 1 Black stones captured by White: 0 result: Black: 23 = 17 territory + 6 prisoners White: 26.5 = 18 territory + 4 prisoners + 4.5 komi Result: W+3.5 --(the content above was generated automatically by GoVariantsTransformer)-- ]B[gk])(;C[This goes further into White’s territory. I play out one sequence; it gives Black winning by 1.5 points. (My endgame technique isn’t very good so this isn’t a definitive judgement.)]B[hb];W[gb];B[hc];W[hd];B[ia];W[id];B[gc];W[fc];B[ha];W[dk];B[ej];W[gk];B[ga];W[fb];B[dd];W[ed];MA[kh][kf][cf][bc]C[move 103 White stones captured by Black: 1 Black stones captured by White: 2 result: Black: 22 = 16 territory + 6 prisoners White: 20.5 = 12 territory + 4 prisoners + 4.5 komi Result: B+1.5 --(the content above was generated automatically by GoVariantsTransformer)-- ]B[gj]))(;C[The variations I give after this atari seem to be the best White can do. All have Black winning by a few points.]B[ie];W[jf];B[if];W[ig];B[ge];W[hh];B[he];W[hg];B[fe];W[id];LB[gb:A][ga:B]C[If black ‘A’ instead, then white ‘B’, and I don’t see how Black can save the stone at ‘A’.]B[hb];W[gb];B[hc];W[hd];B[ha](;W[gc];B[fj](;W[fd];B[ed];W[dk];B[ej];W[dd];B[gk];W[ee];B[ga];MA[kh][cf][bb][eb]C[move 108 White stones captured by Black: 2 Black stones captured by White: 5 result: Black: 24 = 17 territory + 7 prisoners White: 20.5 = 9 territory + 7 prisoners + 4.5 komi Result: B+3.5 --(the content above was generated automatically by GoVariantsTransformer)-- ]W[ed])(;W[dk];B[ej];W[fd];B[ed];W[dd];B[gk];W[ee];B[ga];MA[cf][kh][bb][eb]C[move 108 White stones captured by Black: 2 Black stones captured by White: 5 result: Black: 24 = 17 territory + 7 prisoners White: 20.5 = 9 territory + 7 prisoners + 4.5 komi Result: B+3.5 --(the content above was generated automatically by GoVariantsTransformer)-- ]W[ed]))(;W[ej];B[gc];W[fd];B[fc];W[ed];B[fb];W[ei];B[eh];W[dk];B[dh];W[fj];B[ga];W[gj];MA[kh][cf][bb]C[move 111 White stones captured by Black: 2 Black stones captured by White: 5 result: Black: 20 = 13 territory + 7 prisoners White: 17.5 = 7 territory + 6 prisoners + 4.5 komi Result: B+2.5 --(the content above was generated automatically by GoVariantsTransformer)-- ]B[gi])))(;C[I play out one endgame sequence here that gives Black winning by 1.5 points.]W[ih];B[gf];W[hg];B[ge];W[he];B[fd];W[fc];B[hb];W[hc];B[gb];W[gc];B[gk];W[fk];B[fj];W[ed];B[fe];W[fb];MA[bb][cf][kh][kf]C[move 95 White stones captured by Black: 0 Black stones captured by White: 1 result: Black: 23 = 18 territory + 5 prisoners White: 21.5 = 13 territory + 4 prisoners + 4.5 komi Result: B+1.5 --(the content above was generated automatically by GoVariantsTransformer)-- ]B[dk]))(;W[hk];TR[ci][cj][bj][ck][ca]B[da]))(;C[If Black plays away with some move like this…]B[hi];C[… then this move destroys Black’s position. White captures five stones and revives four dead stones.]W[jk];C[This move doesn’t work, it just lets White capture two more stones.]B[bd];W[ia]))(;B[cj];W[bk];B[aa];C[This would be a catastrophe for Black.]W[jk]))(;B[cj];W[aa];B[bk];W[kk];B[bj];C[White has saved four stones; a good result for White.]W[ka]))(;C[This is probably better.]B[ki]))(;C[White can save the four stones with something like this…]W[db];B[ec];W[ck];C[…but then Black can return to this cut. It feels bad for White.]B[id]))(;C[Maybe I didn’t play this way because of the black ladder-breaker further on in this variation]W[aa];B[ak];W[bk];B[ca];C[This makes a ladder for White…]W[kk];C[… but this ladder breaker is probably good enough for White.]B[kh](;C[I think White needs to answer this way here.]W[kg];C[However now Black can move out; the ladder doesn’t work anymore.]B[aj];W[kj];B[ai];W[bi];C[I don’t think White can capture these stones; I think this is good enough for Black.]B[jh])(;W[ki];LB[jd:A][jf:B]C[With the weaknesses at ‘A’ and ‘B’, I think White’s overall shape is too thin for this to be reasonable for White.]B[kg]))(;C[At the time I didn’t think of this move. Now, I wonder if it would be the best option. I rather like this move, it seems quite flexible. (I only thought of this move on 2021-07-20, after having already published a previous version of this commentary.)]W[kk](;LB[ka:A][ak:B]MA[ib][ic][jc]TR[kb][kk]C[Black coming out of atari here seems like the biggest move. Now the three ‘X’ stones seem to be dead, but maybe White can still win. The connection (between the two ‘△’ stones) is good enough for White; I give two variations to show how White could handle black moves at ‘A’ or ‘B’. ]B[id];C[I show one possible continuation.]W[hd];B[ie];W[jf];B[he];W[ge];B[gd];W[hf];B[hc];W[if];B[hd];LB[ek:A][hj:B]C[I believe this move (aiming to reduce the black framework) is possible for White. Later on White can move this stone out with ‘A’ or ‘B’ or something similar. This is much better for White than in the game.]W[gk])(;B[ka];W[aa];B[ak];W[bk])(;B[ak];W[ka](;B[kj];MA[jb][ja]W[jk])(;B[aj];MA[kk][ka][kb][ab][bb][bc]C[The six marked stones are safe now. I give a few variations showing unsuccessful attempts by Black to capture them.]W[id](;B[kj];W[jk];B[ia];W[jj];B[ji];W[ij](;MA[ik][ia][ja][jb]C[This move or something similar is needed to protect the four ‘X’ stones.]B[gk];W[ii];B[ih];W[jh];B[ki];W[hh];B[ig];W[hg];B[if];W[hf])(;B[ii];W[hk];B[ha];W[ga];B[hb];W[gb];B[hc];W[gd]))(;B[kh];W[kg](;B[jj];W[kj];B[ki];W[jk];B[ia];W[ij];B[ji];W[ii];LB[hk:A]MA[ik][ia][ja][jb]C[This move means that white ‘A’ no longer captures the ‘X’ stones. But…]B[ih];LB[jh:A][bk:B]C[… Black can’t defend against both ‘A’ and ‘B’.]W[ai])(;B[ki];W[jk];B[ia];W[ij];B[jj];W[kj];B[ji];C[Now this is like the previous variation. Black’s shape is too weak for this to work here.]W[ii]))))))(;C[I give a few variations but I haven’t really looked at this very deeply.]W[ad];LB[jd:A][bd:B]C[The cutting points at ‘A’ and ‘B’ are hard to handle for White.]B[ab](;W[dd];B[id];W[hd];B[ie](;W[he];B[if];W[gc];B[ed];W[gg];B[ig];W[fg];C[This feels bad for White.]B[ef])(;W[hf];B[hc];W[hb];B[he];W[gc];B[ge]))(;W[ie];B[bd];W[be];B[cd]))(;C[I reckon this is the best option for White.]W[ke](;C[This doesn’t seem good…]B[ie];C[… because White can move out and capture two stones in a ladder.]W[ab])(;B[ab];C[This seems to be better for White than the game.]W[ie])))(;C[This way leads to an exchange.]W[kc];C[Black cuts…]B[ic];W[id];B[hc];C[… and both sides have captured a stone.]W[ie];C[This seems to be quite a close game to me; I’m not sure if there is an advantage for either side. In any case, with White I prefer this variation to the game.]B[dj]))(;W[ja];B[jc];W[ia];B[ka];LB[ik:A][kb:B][bc:C][ka:D][ba:E]C[I believe this way would be fine for White; it certainly seems better than the game. A few thoughts about this position: 1) The stone at ‘A’ looks a bit useless now; 2) and there isn’t any immediate way for Black to capture the cutting stone at ‘B’. 3) ‘B’ and ‘C’ are a bit weak but so are ‘D’ and ‘E’.]W[jk]))(;C[I don’t really know how Black could play here. Maybe this is feasible. I give two possible continuations.]B[ac](;C[I didn’t think very deeply about this variation. At the end of this variation the result seems promising for Black.]W[ab];B[bb];W[kc];B[ad];W[kd];B[ae];W[ke];B[af];W[bg];B[cd];W[ah];B[kh];W[kk];B[kf];W[jf];B[jh])(;C[I prefer this move.]W[jk];B[ab];W[jj];B[be];C[Perhaps this is good enough for White.]W[bg])))

Revisions
(2021-07-20): Added some analysis for an alternative to white 34.