The game of chess lends itself to narratives of many genres. Some games are epic in scope, the chess equivalent of Lawrence of Arabia or The Godfather. In other instances the odd placement of the pieces gives the game a science fiction feel. Some of my most memorable games—not in a good way—have been tragedies or farces.
I have a special place in my heart for the chess version of the morality play, which goes something like this: I have the initiative, which allows me to pursue a path of righteousness against my foe, the enemy king. I am then tempted to cash out the initiative for a small material gain (greed), for which I am punished by being forced to endure the purgatory of my opponent’s long-lasting initiative, which I must patiently extinguish over many moves.
My penultimate game in the Tuesday Night Marathon was an example of winning through suffering, a skill that I feel uncommonly good at. Something about getting a bad position that I know will eventually rebound makes me more patient, and surviving chess purgatory requires many repeated short but accurate calculations, something that I’m quite capable of even when I’m not in great form overall. Here’s the game:
Hayes Marvin (1876) - Andy Lee (2328), Tuesday Night Marathon (6), 8/13/24
1 d4 d5 2 Nf3 Nf6 3 Bf4 e6 4 e3 c5 5 c3 Nc6 6 Nbd2 cxd4 7 exd4 Nh5:
When you’ve entombed your own fianchettoed bishop for two consecutive rounds, it’s probably time to return to classical opening theory for a round or two. This variation was all the rage against the London System a few years ago. I like how it creates problems with white’s ideal setup from a positionally sound basis. Look at what happens when white goes full caveman and throws all his pieces forward: 8 Bg5 (the main line is 8 Be3 Bd6 9 Ne5 g6 10 g4 Ng7 11 h4) 8 … f6 9 Ne5 g6 10 Bh4 Bd6 (I also considered 10 … Be7, breaking the pin immediately, which happens to be the only move in the Lichess database) 11 Bb5:
The e6-f6-g6 wall of pawns is unconventional, but quite effective. I considered 11 … 00 12 Nxc6 bxc6 13 Bxc6 Ba6 14 Bxa8 Nf4 before deciding that black will have a strong attack without giving up material: 11 … Qb6 12 Bxc6+ (what else?) 12 … bxc6 13 Nd3 Ba6. After the game I wondered if I had played this obvious move too quickly, as there’s a case to be made for 13 … 00 as well. However, there’s nothing wrong with putting the bishop on a6 right away. White continued 14 Nb3 and I had to restrain myself from playing 14 … Bxd3 15 Qxd3 Nf4. If white plays the odd 16 Qf1, the first move I considered, 16 … a5 17 Bxf6 00! is crushing. But 16 Qf3 is far more logical, and white is only slightly worse. I suddenly remembered that tactics always work better when the king is safe and all the pieces are developed, so 14 … 00 15 Nbc5:
Why didn’t white castle when he had the chance? If 15 00 I was planning 15 … Qb5 16 Nbc1 e5, when white’s position is starting to crack. I’d already planned what I was going to do after either knight hop to c5, otherwise I might have preferred 15 … Bxd3 16 Nxd3 e5 from the position above. Instead I double-checked my tactics and played 15 … Bxc5 16 Nxc5 Qxb2! I have to admit, I hadn’t been this happy all tournament long. I never feel quite right if I haven’t sacrificed a piece recently, even if the material will be recovered in short order. If 17 Nxa6 Qxc3+ 18 Kf1 Qc4+ (I also looked at 18 … Nf4 19 Nc5 Qc4+ 20 Kg1 Ne2+ 21 Kf1 and there’s no obvious win) 19 Qe2 Qxd4 forks the Ra1 and Bh4:
Even after winning back the e-pawn, white’s position will be quite horrible. My opponent took his time and found a much more interesting resource: 17 Rb1! Qxc3+ 18 Qd2 Qxd2+ (18 … Qa3? 19 Rb3 or 18 … Qc4? 19 Rb4 and the queen is trapped) 19 Kxd2:
The elements of the morality play are in order. I had the initiative and greedily traded it for two of white’s queenside pawns. In exchange, white’s rooks are connected and can assume active posts. The knight on c5 is a monster, particularly after my bishop retreats to c8 to defend the e-pawn. Now the suffering will begin.
Of course I wanted to play 19 .. Bc4, but it’s too early to give back a pawn, and the bishop does little on the a6-f1 diagonal. After 19 … Bc8 it’s at least also covering the b7 square to keep out the white rook. There followed 20 Rhe1 g5 21 Bg3 Nxg3 22 hxg3:
I admit that it looks like things are even worse than they were a few moves ago. Trading the black knight for the white bishop accentuates the dominant status of white’s remaining minor piece, but it also accomplishes a few key things: first, there will be no trade of knight for knight that might transition to an opposite-colored bishop ending. The white bishop, now deceased, will not be allowed to station itself on the h2-b8 diagonal; white will have to keep a rook on the b-file, lest black activate with a future Rb8. And finally, bishops are just much better than knights in rook+minor piece endings. Even a terrible bishop like to one on c8 can be activated with a pawn sacrifice, and black has two pawns to give.
In the position above the computer thinks that black can break the position open immediately: 22 … e5!? 23 dxe5 fxe5 24 Rxe5 Rxf2+ 25 Ke3 Rf5. I was concerned here about 26 Re8+, pinning by bishop on the back rank, but the machine sees all the tactics: 26 … Kf7 27 Rd8 Ke7 28 Rh7 d4+!:
The point is that the only way to guard the knight, 29 Kxd4 Rd5+ 30 Kc4, runs into 30 … Ba6+, picking up the exchange.
That’s all very clever, but it goes against my philosophy of how to play these kinds of positions: black’s position is bad, there will be further suffering, but the black pieces have more potential for future activity than their white counterparts. There’s no need to rush to open things up; after all, which of white’s pieces can be improved? The game continued 22 … Re8 23 f4 h6 24 a4 Kf7 25 Re3 Rg8:
Black is threatening to come crashing through on the g-file, so white has to take immediate action. Neither of us noticed white’s best plan, 26 Nxe6! Bxe6 27 Rb7+ Kg6 28 Rxe6—black is still better here, but not immediately winning. Instead white rolled the dice with 26 f5?, perhaps hoping for 26 … e5? 27 g4!, truly shutting down the bishop. I continued to follow the law of greed: 26 … exf5 27 Rbe1 and now I thought that 27 … f4 was basically forced to free some squares for the bishop. I was worried about 27 … Kg6 28 Re8 Rxe8 30 Rxe8:
White’s pieces dominate black’s, but according to the machine 30 … Kh5 is comfortably winning. White has no way to make further progress and must give ground to avoid losing his remaining kingside pawns.
Fortunately, my choice was also quite strong, and the game continued 28 Re7+ Kg6 29 Rc7 (a strange choice; white should have first captured on f4) 29 … fxg3 (greed forever!) 30 Rxc6:
The black pawn g3 is a huge problem; it gives black all kinds of ways to win the position. As it is, after 30 … Bf5! white can safely resign, but instead I played carelessly: 30 … Rb8?, missing 31 Nd3! entirely. Now there’s no entry point for the rook on b2 and the black king doesn’t look so safe anymore. I burned a bunch of my remaining time and finally settled on 31 … Rb7. I wanted to play 31 … Bf5 instead, but I overestimated white’s attacking chances on the seventh rank. After 32 Ne5+ Kg7 (not 32 … Kh5? 33 Rh1+ and mates) 33 Rc7+ Kh8 34 Nf7+ Kh7 the discovered checks don’t amount to much: for example, 35 Nxg5++ Kg6 36 Nf3 Be4 and black wins.
The suffering continued for a while longer: 32 Rec1 Bd7 (I was still convinced that I had to avoid checks on the seventh) 33 Rd6 (33 Ne5+ Kg7 34 Rc7 Rxc7 35 Rxc7 fxe5 36 Rxd7+ Kf6 and black will win) 33 … Bf5 34 Rf1 g4 (I thought this was yet another forced move, but it turns out that 34 … Bxd3 35 Rfxf6+ Kh5! 36 Rxh6 Kg4 37 Kxd3 Rb2 also wins, as black picks up the g-pawn) 35 Nf4+ Kg5 36 Rxd5 Rb2+ (the only winning move) 37 Kc3 Rc2+ 38 Kb4 Rb8+ 39 Ka5 Rf2!:
The suffering comes to an end, at last. White played 40 Ne6+ Kg6, but now he’s out of checks: 41 Nf4+ Kf7! 42 Rxf2 gxf2 43 Rxf5 f1=Q and it’s over. He played a few more moves before throwing in the towel: 41 Re1 Bxe6 (good riddance to the annoying white knight) 42 Rxe6 Rxg2 43 Re4 h5 44 Re1 Rf2 45 Rc5 g2 46 Rg1 h4 47 Rc3 Rb6 (weaving a mating net) 48 d5 Rd2 and resigns.
Not a perfect game by any means, but I avoided most of the pitfalls in the ending. Going over my calculations from the game with the computer was rather eye-opening: as I’ve tried to note in the annotations above, there were often multiple acceptable moves in positions where I thought that my moves were basically forced. It’s a good reminder to always try to look a move or two deeper and not abandon variation simply because they are superficially bad.
Next week is the final round. Second place in the tournament is very much still on the table, and then I’ll likely resume my chess hibernation for another year.