Game #466: the London System vs Nc6, and my biggest upset
On May 20, 2024, in game #466 of my rapid journey, I beat bobbo6767 (1286) with White while rated 1035. It took twelve moves and six and a half minutes. Black played an early knight to c6 against my London System, and the London System vs Nc6 position gave me the one thing I needed: a target. My rating went to 1048, a gap of 238 points to my opponent, measured the way the series measures every upset (my rating after the game against theirs). It is the biggest upset in my first 2,116 games.
That number still surprises me. Before this game I had played thirteen games against opponents rated 1200 or more: twelve losses and one draw. Across the whole journey, against players 200 or more points above me, this is my only win in 17 games.
Where I was that day
This game sits in week 20 of the journey, and it was the second game of a strange Monday. Game #464, the night before, was a loss to bshtylla (1254). That morning, in game #465, I lost to DewaCatur_19 (1274) and left the game in a lost endgame. So I came into this one on two losses against players rated over 1250, and the matchmaking gave me a third.
| Game | #466, May 20, 2024 |
| Opponent | bobbo6767 (1286) |
| My colour, rating after | White, 1048 (from 1035) |
| Time control | 20 minutes, no increment |
| Opening | Queen's Pawn Opening, Zukertort Chigorin Variation (D02) |
| Result | Win, Black resigned after move 12 |
| Accuracy (Stockfish) | Me 92.6, Black 76.2 |
| My blunders / mistakes / inaccuracies | 0 / 0 / 1 |
Chess.com files it under the Zukertort Chigorin name, but what I actually played was my usual London setup: pawn to d4, bishop out to f4 early, pawns on e3 and c3, knights on f3 and d2.
The opening: Black's knight comes to c6
The moves were 1.d4 d5 2.Bf4 Nf6 3.e3 Nc6 4.Nf3 Bf5. Black copied my bishop and put the queen's knight in front of the c-pawn, which blocks the c-pawn from going to c5 later.
I played the knight to d2, which the engine didn't mark as a mistake, but it wasn't its first choice either. Stockfish wanted the bishop on b5 straight away, pinning the knight. When I checked the position again for this post, it rated the bishop move "slightly better for white" and my knight move a bit lower. I only found the idea two moves later.
After 5...e6 6.c3 h6 I did play 7.Bb5. Black answered with 7...g5, a pawn push on the side where Black's own king still stood, and I moved my bishop from f4 to e5 to stay out of its way.
Moment 1: the pawn push that gave me a target
On move 8 Black pushed the a-pawn to a6, attacking my bishop on b5. The engine called it a mistake: the evaluation went from +0.62 to +2.14 (evaluations are from White's side, so plus means better for me).
I took the knight with check, and Black took back with the b-pawn. Now the c-file had two black pawns, one on c6 and one on c7, and the one on c6 had nothing behind it. The engine's line from here was exactly what I played next: queen to a4, attacking that pawn.
I'd like to say I saw all of that. The honest version is simpler: the bishop was attacked, and taking the knight with check looked cleaner than retreating. The engine agreed, for better reasons.
Moment 2: the rook move that lost the game
After 10.Qa4, Black had to deal with two problems at once. My queen was aiming at c6, and my bishop on e5 was hitting the knight on f6, with the rook on h8 behind it on the same diagonal.
Black played the rook to b8, attacking my pawn on b2. The engine called it a blunder: +1.57 to +8.84. Its choice was bishop to d6, challenging my bishop on e5, which it scored only "slightly better for white". The rook move did nothing about either threat.
Moment 3: my one inaccuracy
I took on c6 with check. The engine preferred bishop takes knight on f6: a double attack (one move that threatens two things at once) on the queen on d8 and the undefended rook on h8. When I checked it again, Stockfish rated the bishop capture "completely winning" and my queen check "winning", so I gave away a bit of the edge and nothing more. The game analysis logged it as my only inaccuracy.
Moment 4: the double attack that ended it
Black's king stepped up to e7, out of the check. That left the c7 pawn without a friend.
The bishop took on c7, attacking the rook on b8 and the queen on d8 at the same time. Black resigned right there. When I put the final position back into the engine, every reply it found for Black was still "completely winning" for me.
I had about sixteen minutes left on my clock. The whole game, from first move to resignation, lasted from 15:24 to 15:31.
What the engine says about game #466
All twelve moves were in the opening phase, at 92.6 accuracy for me. One inaccuracy, no mistakes, no blunders. Black had one mistake (8...a6) and one blunder (10...Rb8). It wasn't a brilliant game. It was a short one where the stronger player made two bad moves in a row and I didn't drop anything back.
The London System vs Nc6, from this game
I don't want to turn one twelve-move game into an opening theory lesson. But a few things from it are worth keeping:
- An early knight on c6 can be pinned. The bishop to b5 was the engine's move on move 5. When the knight is pinned to the king, it's a piece that can be taken with check later.
- Doubled pawns on the c-file are targets. After bishop takes c6 and the b-pawn takes back, the pawn on c6 stood alone. My queen on a4 hit it right away.
- Look for the capture that hits two things. On move 11 the stronger capture attacked the queen and a rook. On move 12 my bishop did attack both, and that was the end.
- Rating gaps aren't verdicts. My opponent was 251 points above me when the game started. My opponent still had to find moves, and on move 10 didn't.
What came next
The upset didn't start a run. That same evening I lost to Marleon777 (1304) and then to stodra (1329), the second with 93.3 accuracy. Those were the first two of six losses in a row, and I finished the week at 1026. The story of that streak is in week 20; the week before, 1022 → 1038 in 16 games, is week 19. The whole series is on the journey page, and you can replay game #466 on Chess.com.
If you want your own games read this way, with the moment, the better move and the reason in plain words, that's what Chess Coach does inside Claude or ChatGPT. If engine numbers like +8.84 feel like noise, I wrote a guide on how to read Stockfish analysis, and there's more on how the analysis works.
Frequently asked questions
How do you play the London System against an early Nc6?
In my game the knight on c6 became the target. Stockfish wanted my bishop on b5 to pin it (a pin is when a piece can't move without exposing something more valuable behind it). I played that bishop move two moves later, took the knight with check, and my queen on a4 went after the pawns on c6 and c7.
How did a 1048 beat a 1286 in 12 moves?
Black made two bad moves in a row by the engine's count: 8...a6 (a mistake) and 10...Rb8 (a blunder). The rook move left the pawn on c6 to my queen with check, and my bishop then took on c7 attacking the rook and the queen at once. Black resigned.
Did you play game #466 perfectly?
No, but close. Stockfish gave me 92.6 accuracy with no blunders and no mistakes, and one inaccuracy: on move 11 bishop takes knight on f6 was stronger than my queen check on c6. Both were winning.