Nimzo-Indian Defense: A wild ride

In two days, barring a resumption of the war with Iran, my students will be returning to school. The free time I enjoyed from not having homework to check and tests to grade, as well as the surplus energy from not having eight classes to manage every day, will slowly fade into memory as the teacher grind takes hold again.

In today's unrated game Lichess matched me with one mightypidgeon, provisionally rated 1618 Blitz. Our opening developed into what Lichess classifies as the Noa Variation of the Classical Nimzo-Indian, and I emerged from it with a comfortable advantage. The middlegame and engame, however, were a roller-coaster ride of eval swings: ↗+3.2 ↘-3.2 ↗+5.0 ↘-4.4 ↗+4.2 ↘#-5, ultimately ending in me winning by checkmate.

---

Position 1: White to move

Assigned the white pieces, I opened 1.d4, and after the moves 1...Nf6 2.c4 e6 3.Nc3 Bb4 4.Qc2 we were in the Classical Variation of the Nimzo-Indian Defense. Here as Black I typically attack White's d4-pawn with 4...c5, which is evidently known as the Berlin Variation; but opponent in this game opted for 4...d5, the Noa Variation. 5.cxd5 was correct, but 5...Nxd5 less so. I correctly attacked with 6.e4, and Black moved his knight a third time to exchange with 6...Nxc3 7.bxc3. Black's bish, under attack from my c-pawn, retreated with 7...Be7, whereupon the engine wanted 8.Bb5+ to induce 8...Bd7 and then retreat to e2 or d3. Instead I continued 8.Bd3, in anticipation of Black's 8...0-0.

Here I was already better, with 9.e5 the obvious next move and an attack on h7. After Black blocked with 9...g6 the engine insists on 10.Bh6 or offering a pawn sac with 10.h4, but I continued 10.Nf3. Black needed to attack my center with 10...c5 but instead played the slow 10...b6. Here 11.Be4 was the best move, but it's reassuring to know the engine approves of my 11.h4 despite not being castled. Black fianchettoed with 11...Bb6; and again the engine demanded 12.Bh6 to follow up with 13.h5 and potentially chop on g6 and after 14...hxg6 sac the bish with 15.Bxg6. I think the idea behind my 12.Ng5 was to follow up with 13.f4 and 14.h5, but 12...Bxg1 surprised me and the g-file's opening prompted a change of plans. That plan involved summoning the rook from a1 to g1. 13.Rh2 was the wrong way to kick Black's bish as the engine wanted 13.Rg1, but worse than that was the move I played after Black retreated his bish with 13...Bb7: my advantage of +1.7 fell to -0.5 when I developed my bish with 14.Be3.

1. Why did I choose the move 14.Be3?
A path needed to be cleared in order for my a1-rook to come to the kingside, and that involved developing the dark-square bish. On d2 it would be passive, and on f4 it would get in the way of my f-pawn.

2. Why is my move not ideal?
14.Be3 invited Black's 14...c5, which started chipping away at my center.

3. Why is the better move better than my chosen move?
The engine's top choice was 14.Qe2, a hard move for me to find as it broke up my queen-and-bish battery in favor of a more tactical idea. That tactic involved pushing the h-pawn and bringing the queen to g4 or h5. White's attack is dangerous enough that castling can be delayed, and if still 14...c5 then White sacs two pieces, e.g. 15.Nxh7 Kxh7 16.Qh5 Kg8 17.Bxg6 fxg6 18.Qxg6+ and Black's king is fully exposed.

---

Position 2: White to move

After Black correctly struck at my center with 14...c5, the engine wanted castles with 15.0-0-0 or offering a bish trade with 15.Be4. My 15.Kd2, to clear a path for my a1-rook, was a blunder. Black correctly chopped with 15...cxd4 but after 16.cxd4 developed his knight awkwardly with 16...Na6. My 17.a3 to take away b4 from Black's knight was the top move, as were 17...Rc8 and 18.Qa4. Black should have threatened a queen-fork with 18...Qc7 but retreated his knight with 18...Nc7.

Proceeding with my plan, 19.Rg1 mobilized a fifth attacking piece, but this was a blunder because after 19...Nd5 and 20...b5 Black's attack is faster than mine. What the engine wanted was what I played on the next move: 20.Nxh7. Black should have declined my sac in favor of 20...Nxe3 as 20...Kxe3 swung the eval bar from -3.7 to +4.6. I correctly continued 21.h5, and after Black's 21...f5 my 22.hxg6+ was also correct. Black's king sought refuge behind my g-pawn with 22...Kg7, allowing 23.Rh7+. I didn't play that, however, and my 23.Bh6+ flattened the eval bar from +3.5 to 0.0. It got worse, though, as after 23...Kg8 and 24.Bxf8 Black's advantage soared to -4.4.

1. Why did I choose the move 24.Bxf8?
24.Bxf8 won back material and cleared the h-file for my rook.

2. Why is my move not ideal?
My bish was one of two pieces controlling the g5-square, whose importance I failed to appreciate. 24.Bxf8 allowed Black to launch his attack with 24...Bg5+. After 25.Ke1 to avoid getting royal-forked Black's 25...Rc1+ is effectively game over.

3. Why is the better move better than my chosen move?
The engine's top move was 24.Qd1. It is telling to note that this is the move I played on my next turn. I understood the need for another piece in my attack, but I acted on that understanding too late. 24.Qd1 needed to be played at that moment because Black's 24...Nc3 would open the d-file for his queen and the a8-h1 diag for his bish and without setting up major threats of my own Black would swarm my king first.

---

Position 3: White to move

Black missed 24...Bg5+ and played the more instinctive 24...Bxf8. I was already down -2.5 on the eval bar, but 25.Qd1 sunk me to -5.1. Black's 25...Rc7 boosted me to +4.8 on account of allowing 26.Qh5, and 26...Bg7 invited 27.Qh7+ Kf8. After 28.Rgh1 the eval bar flipped from +4.0 to -#5.

1. Why did I choose 28.Rgh1?
I saw a sequence where with my rooks stacked on the h-file I would either checkmate Black or win his queen.

2. Why is my move not ideal?
I failed to anticipate 28...Qg5+ and my king about to get mated.

3. Why is the better move better than my chosen move?
Tactics. The engine's top moves were 28.Qh8+ Bxh8 29.Rxh8+. If 29...Kg7 Black's queen is lost; if 29...Ke7 then 30.Rh7+ and after 30...Ke8 or 30...Kf8 White advances the g-pawn, then Black has to sac, and when all is said and done White is left with a bish and rook v. a bish and knight.