Game #442: the Accelerated London System with c4, a mate in four I missed, and a loss to a 1385
On May 12, 2024, in game #442 of my rapid journey, I played the Accelerated London System with c4 against onehorseopen (1385), rated 1011 myself. I was lost by move 9, back to winning by move 16, and on move 28 the engine found a forced mate in five for me. I played the first move of it and not the second. Fifteen moves later, with 2 minutes 44 seconds on my clock, I walked into a one-move checkmate.
The series picked this game as an Accelerated London lesson. The opening part is about one pawn capture on move 8. The rest is about what happens when the opponent is 374 points higher and I start playing on my clock instead of the board.
Where I was
Game #442 sits in week 18 of the journey, a week I finished 1017 → 1022 with 6 wins, 10 losses and 2 draws. Sunday, May 12 was a ten-game day, and this was the fourth game of it. I had lost five games in a row (#436 to #440), then beaten amirhossein1980 (1136) in game #441. Game #442 started less than five minutes after that win ended.
My opponent was rated 1385. That number matters to this series: it's the peak rating I reached, in November 2025, 1,430 games after this one. At the time, only one opponent I'd faced had a higher rating, Chandu2898 (1405) four days earlier, and I lost that one too.
| Game | #442, May 12, 2024 |
| Opponent | onehorseopen (1385) |
| My colour, rating after | White, 1009 (from 1011) |
| Time control | 20 minutes, no increment |
| Opening | Queen's Pawn Opening, Accelerated London System (D00) |
| Result | Loss, checkmated on move 43 |
| Accuracy (Stockfish) | Me 68.3, Black 57.2 |
| My blunders / mistakes / inaccuracies | 5 / 3 / 5 |
The loss cost me two rating points.
The Accelerated London System with c4
The moves were 1.d4 Nf6 2.Bf4 d5 3.c4. Bringing the bishop to f4 on move 2, before the knight, is the "Accelerated" London. Pushing the c-pawn two squares on move 3 is the more aggressive version: it hits Black's d5 pawn straight away, like a Queen's Gambit. Stockfish preferred the knight to f3 on move 3, but rated c4 only a quarter of a pawn worse and didn't mark it.
Black played the bishop to f5 and the knight to c6, and on move 5 the knight jumped to b4. From b4 it attacked my a2 pawn and the c2 square next to my king. When I re-checked the position for this post, the engine's best answer was the one I played: queen to a4 with check, which also attacked the knight on b4. It scored that at +3.14. Black blocked with the knight back to c6, which the engine marked as an inaccuracy. So far the c4 version had done what I wanted: Black's pieces were going backwards.
Moment 1: the pawn I shouldn't have taken
On move 7 I put my knight on e5 (an inaccuracy; the engine wanted the other knight out to c3), and Black put the bishop on d7, behind the knight on c6.
I took on d5 with the pawn, attacking the knight on c6. It looked like I was winning a piece. But the knight on c6 didn't have to run: it took my knight on e5, and moving off c6 uncovered Black's bishop on d7, which now attacked my queen on a4. The evaluation dropped from +0.89 to -3.59 (evaluations are from White's side). The engine's move was the simple trade first: my knight takes the bishop on d7, Black takes back, and the position is "slightly better for white".
Then I made it worse. On move 9 I pinned with the bishop on b5, and the engine called it a mistake: -7.31. It wanted the queen on b3, hitting the b7 pawn. In two moves I'd gone from a small edge to a lost position against a player rated almost 400 points above me.
Moment 2: the right capture, one move too late
Black didn't convert it. On move 12 Black's rook went to a7, a blunder that took the evaluation back to +0.72, and after my knight took a pawn on b5, Black's rook went to b7, another blunder. Now the engine had me at +2.83.
The knight takes on c7 with check was winning. I took the pawn on a6 with my queen instead, and the engine dropped to -2.93. Then Black blundered again, putting the queen on c8, and it was +1.44 for me.
Moment 3: the same capture, now wrong
Here's the moment I find most useful from the whole game. The knight capture on c7 was available again.
One move earlier it was the best move on the board. Now it was a blunder: -5.53. The difference is Black's queen on c8. If Black's rook takes my knight on c7, the rook leaves b7 and opens the diagonal from c8 to my queen on a6. The engine's line shows Black winning material that way. I had learned "knight takes on c7 with check" as a good idea one move before, and I played it again without asking what had changed.
Black didn't take with the rook. The king went to d8, the third blunder in a row in this sequence, and the engine jumped to +6.95. My rook came to c1, my knight checked from e6, and when Black took it, my rook took Black's queen on c8. By move 22 I had queen and rook against two rooks, knight and bishop, and the engine scored it between +7 and +9.
Moment 4: mate in four, and a queen check instead
On move 28 Black's rook went back to a8, and the engine found a forced mate for me: mate in five. I played the first move, rook to d1 with check, and Black's king went to e8.
The engine's mate started with a quiet pawn push to e6, threatening mate on the next move. I gave a check instead. The check attacked two pieces, and it kept me "completely winning" at +6.95, so the analysis didn't even mark it. But the mate was gone. I had 7 minutes 35 seconds on my clock. My opponent had more than 14.
Moment 5: four minutes left, and the pawn push
For the next few moves I chased Black's king with queen checks and picked up the h5 pawn, while Black took my pawns on b4 and a5. On move 36 I pushed my f-pawn to f5 (a mistake, +6.00 to +3.45), and on move 37 I pushed it again.
The f6 push threatened mate, and I'm sure that's why I played it. Black just took it with the e-pawn, and the advantage went from +3.20 to +0.69. The game analysis preferred e6; when I re-checked, the engine put queen to g4 first (+3.54), with e6 close behind (+3.24). I had 4 minutes 31 seconds left. On the next move I moved my queen to f5 instead of g4, and the engine had Black better for the first time since move 15.
Moment 6: the checkmate
My queen move attacked three of Black's pieces at once. It also let Black's rook go to f5 with check, and with the rook on e8 covering the e-file and the bishop on c5 covering f2 and g1, that was checkmate. From b7, my queen could have dropped back to f3 to block a check on the f-file; from b5 it couldn't. The engine's move, g4, takes the f5 square away from the rook. It would still have been "black is winning", but not mate. I had 2 minutes 44 seconds left. My opponent had over 12 minutes.
What the engine says
My accuracy was 73.8 in the opening and 66.6 in the middlegame; the game never reached an endgame by the analysis's count. My best winning chance during the game was 97.5 %. Stockfish counted 5 blunders for me and 4 for Black, and the clock tells the rest: I spent most of my time before move 30, and played the last eight moves with under five minutes.
The lesson I took from game #442
The opening lesson about the Accelerated London System with c4: the pawn on c4 invites tactics on both sides. When Black's knight goes to b4, queen to a4 with check is a strong answer. But on move 8, before taking on d5, I should have traded off the bishop on d7 first, because my knight on e5 was the piece that would hang once Black's knight was free to move.
The bigger lesson is about Moment 3. A capture that was the best move one turn ago can be a blunder now, because the opponent's last move changed one line. I now try to ask, before playing a move I had already planned: what did the opponent's last move attack, and what did it open?
What came next
I lost my next game too, #443 against LINK8x8 (1016), on time in a position the engine scored as winning for me. Later that day I won three in a row, the first of them an Accelerated London against pop-meister (1093), and finished the week at 1022. The previous Queen's Pawn lesson is game #412, where my back rank nearly cost me, the week is in the week 18 diary, and every week and milestone is on the journey page. You can replay game #442 on Chess.com.
If you want your own games checked like this, every swing found and explained in plain words, that's how Chess Coach's analysis works. And if you've wondered why the built-in review doesn't explain moves like 15.Nxc7+ here, I wrote about Chess.com's game review not explaining moves.
Frequently asked questions
Is c4 a good idea in the Accelerated London System?
In my game, 1.d4 Nf6 2.Bf4 d5 3.c4 was rated a little worse than 3.Nf3 by Stockfish, but not an inaccuracy. The trouble came two moves later: Black's knight jumped to b4, aiming at c2, and my queen check from a4 was the engine's best answer, scored +3.14 when I re-checked it.
What was the opening mistake in game #442?
8.cxd5. My knight on e5 could take Black's bishop on d7, which the engine preferred (+0.86). Instead I took the pawn, Black's knight took mine on e5, and the evaluation fell to -3.59. My next move, bishop to b5, made it -7.31.
How did you lose a winning position against a 1385?
After move 28 the engine found a forced mate for me in five moves. I played the first move of it, then chose a queen check instead of the pawn push that mated in four. I kept a big advantage for a while, but with under five minutes on my clock three pawn and queen moves gave it back, and I was checkmated on move 43.