Game #412: the London System vs d6, and the back rank I forgot
On April 29, 2024, in game #412 of my rapid journey, I played the London System vs d6 against krzysztofmielczarek (1144), rated 997 myself, and won in 42 moves. The opening went almost perfectly: 95.8 accuracy by Stockfish's count, and a +3 position by move 12. Then, on move 25, I put my rook on the seventh rank and left Black a mate in one on my back rank. My opponent didn't see it, and I didn't either.
The series picked this game as a Queen's Pawn Opening lesson. It teaches two things, and only one of them is about the opening.
Where I was
Game #412 was the first game of week 17 of the journey, the week I climbed back above 1000 (997 → 1017, 9 wins and 10 losses). The night before, I had lost game #411 to bmw2228 (1008) with 57.6 accuracy. This win took me to 1008. Thirteen minutes after it ended, I started game #413 against zafpar (1191), the highest-rated player I had beaten up to then.
| Game | #412, April 29, 2024 |
| Opponent | krzysztofmielczarek (1144) |
| My colour, rating after | White, 1008 (from 997) |
| Time control | 20 minutes, no increment |
| Opening | Queen's Pawn Opening, 1...d6 2.Bf4 (A41) |
| Result | Win, Black resigned after move 42 |
| Accuracy (Stockfish) | Me 68.4, Black 63.4 |
| My blunders / mistakes / inaccuracies | 6 / 5 / 4 |
The London System vs d6: what I actually played
I opened 1.d4 and Black answered 1...d6 instead of the usual d5. I played my London bishop to f4 anyway, and Black hit the center at once with 2...c5. I propped it up with 3.c3, Black took on d4, and I took back with the c-pawn. My pawn stood on d4, and on move 5 I put a second one beside it on e4. With Black's pawn on d6 instead of d5, nothing stopped it.
The rest was normal development: knights to f3 and c3, bishop to e2, castle. Black castled, played h6 and a6, and put a knight on c6. Through twelve moves the engine marked nothing as an inaccuracy for me, and that's the part of the London System vs 1...d6 I'd repeat: let Black's c-pawn take on d4, recapture towards the center, and take the e4 square while it's free.
Moment 1: the pawn that pushed the knight to the edge
On move 11 I pushed my d-pawn to d5, attacking the knight on c6. The engine's choice was a rook move, but d5 cost only a little. The real swing came from Black's answer.
The knight went to a5, and the evaluation jumped from +0.68 to +2.97 (evaluations are from White's side, so plus is good for me). The reason is that a knight on the edge has few squares. My next move, b4, attacked it, and the engine's line shows Black losing it: Black takes on d5, I take the knight with the b-pawn. I did play b4, the engine's first choice. Black found a way out, queen to c7 and then the knight to c4.
Moment 2: trading my good piece for the knight
I took the knight with my bishop, which felt natural: the piece that had been in trouble for three moves was finally gone. The engine called it a mistake (+2.33 to +0.96). Black's queen took back on c4, attacking my rook on f1 down the diagonal my bishop had just left. The engine wanted the pawn capture on e6 first, keeping my bishop. When I compare the two, the difference is that my move solved Black's problem for them: the badly placed knight was traded for my bishop instead of staying a target.
On the next two moves both of us blundered. Black took back on e6 with the bishop, and I took the d6 pawn with my bishop when a knight move was much stronger. I wrote about that exchange in the week diary. By move 17 the game was level.
Moment 3: the rook on the seventh and the empty back rank
By move 25 the board had thinned out. I had a queen, rook and knight; Black had queen, rook and knight. My king sat on g1 behind pawns on g2 and h2, and neither of them had moved. Black's queen had just landed on d2, inside my position.
The rook to c7 attacked the pawn on f7, and my knight on e5 was attacking it too. I liked the move. But my rook was the only piece guarding the first rank. Once it left, Black had queen to e1, checkmate: the queen comes in from d2, my king can't go to f1, f2 or h1 because the queen covers them, and g2 and h2 are blocked by my own pawns. That's a back-rank mate: the king is trapped on the first rank behind its own pawns. The engine scored the position after my move as mate in one for Black.
Black took my pawn on f4 instead. I had 11 minutes 46 seconds on my clock when I played Rc7, so I can't blame time.
Moment 4: the check that gave everything back
The back rank kept haunting the game. Two moves later, after more blunders on both sides, the engine scored the position as level. My knight on d7 attacked Black's knight on f6, and I took it with check.
The capture lost to a simple pawn recapture on f6. With my knight gone from d7, the d-file was open: the engine's line has Black's rook landing on d1 with check, the same empty back rank, and it rated the position -11.68. My opponent took back with the queen instead, my queen took Black's queen, and suddenly it was a rook endgame with even material. The game's analysis marks my queen recapture as the good move of the game. It was the only one available.
Moment 5: the endgame, won twice
The rook endgame was not tidy either. On move 37, with my rook on b6 and my pawns on a5 and b4, I pushed my h-pawn.
The evaluation went from +5.15 to +0.55. The engine wanted the queenside pawns moving. Over the next two moves we traded errors again, and on move 39 I took the a6 pawn and then f6 and h6. On move 41 Black's rook took my a5 pawn, with Black's king on d5 on the same line. My rook checked from h5, and the rook on a5 was lost behind the king. That kind of attack, through one piece to the one behind it, is a skewer. Black resigned.
What the engine says
My accuracy by phase was 95.8 in the opening, 56.5 in the middlegame and 68.2 in the endgame. All six of my blunders came after the opening. Black made eight. The engine's tally for the whole game is a list of chances both of us passed on.
The lesson I took from game #412
The opening lesson is short. Against 1...d6 and an early c5, recapturing on d4 with the c-pawn and then playing e4 gave me space and easy development. And when Black's knight goes to the edge, a pawn attack can leave it with no squares.
The bigger lesson is the back rank. In this game the same weakness appeared on move 25 (mate in one) and move 27 (a rook check that the engine scored at -11.68). Both times I was attacking something on Black's side. Now, when my king sits behind three unmoved pawns, I check one thing before moving the rook off the first rank: can a queen or a rook land there with check?
What came next
The next game, #413, lasted seven moves and was my first win against anyone rated above 1173: beating a 1191 in the Accelerated London. The one after that I lost with Black to bimalhembram (1035). The whole week is in the week 17 diary, and every week and milestone is on the journey page. You can replay game #412 on Chess.com.
If you want your own games checked for moments like move 25, the threat found and explained in plain words, that's how Chess Coach's analysis works.
Frequently asked questions
How did the London System vs 1...d6 go in this game?
Very well for twelve moves. After 1.d4 d6 2.Bf4 c5 I played c3, took back on d4 with the c-pawn and built a big center with e4. Stockfish gave me 95.8 accuracy in the opening, and when I pushed d5 on move 11, Black's knight jump to a5 was rated a blunder: the evaluation went from +0.68 to +2.97.
What was the worst move of game #412?
25.Rc7. My rook went to the seventh rank to attack the f7 pawn, but my own back rank was empty, and Black had queen to e1, checkmate. The engine scored the position after my move as mate in one. My opponent took a pawn on f4 instead.
Did you win game #412?
Yes. After a run of blunders by both of us, the queens came off on move 28 and I won the rook endgame. On move 42 a rook check on the fifth rank hit Black's king with Black's rook behind it, and my opponent resigned. Stockfish counted six blunders for me and eight for Black.