Game #284: a London System win that ended as a rook vs pawn endgame draw
On March 8, 2024, in game #284 of my rapid journey, I played the London System against hosseinkameli76 (1054) and reached a rook vs pawn endgame that the engine scored as winning for me. It finished as a draw: king against king, insufficient material. I had more than nine minutes on my clock when I made the move that threw it away.
This is one of the games the series picked as a London System lesson, and it earns the label twice: once for the opening, which went better than almost any London I played that spring, and once for everything after it.
Where I was
Game #284 sits in week 9 of the journey, the first week I finished above 1000: 40 games, 18 wins, 19 losses, 3 draws, rating 982 → 1023. I came into this game at 1041 after losing the previous one to CucakYellow (1227). The draw left me at 1041, and I lost the next two games.
| Game | #284, March 8, 2024 |
| Opponent | hosseinkameli76 (1054) |
| My colour, rating after | White, 1041 (unchanged) |
| Time control | 20 minutes, no increment |
| Opening | London System (D02) |
| Result | Draw, insufficient material, move 59 |
| Accuracy (Stockfish) | Me 77.3, Black 77.9 |
| My blunders / mistakes / inaccuracies | 6 / 2 / 4 |
The opening: a bishop with nowhere to go
The moves were 1.d4 d5 2.Bf4 Nf6 3.Nf3 Bg4 4.Nbd2 e6 5.e3 Bb4 6.c3, and Black's bishop backed off to a5 and then b6 when my queen came to b3. On move 8 I pushed my a-pawn to a4, and Black answered a6, which the engine marked as a blunder. Here is why.
The pawn hit the bishop, and its only safe square was a7. With the bishop gone from b6, my queen on b3 had a clear path down the b-file, and it took the pawn on b7 on move 10 and another pawn on a6 on move 11. By move 13 the engine had me at +5 or more. My opening accuracy for the game was 93.2, the best phase I had.
Moment 1: the pawn on c7
On move 17 my queen sat on c6 and my bishop on f4 looked at the pawn on c7. Taking it attacked Black's rook on b8 and queen on d8 at once. The analysis tool describes it as a double attack, a move that threatens two things, and I'm sure that's what I liked about it.
The answer was 17...Qc8. Black's queen attacked my bishop on c7 with my own queen standing behind it on the same file, and Black's rook was free to take my pawn on b2 next. The evaluation fell from +5.51 to +1.31 (evaluations are from White's side). The engine's move was a simple b4, keeping my pawns together and castling next.
Moment 2: the knight that opened the second rank
Two moves later Black's rook was on b2, deep in my position, and my knight on d2 was the only piece between it and my king's side of the board. I retreated the knight to b1.
With the knight gone, the rook could come to e2 with check. The engine scored my move at -4.74. Black didn't find the check and played the bishop to b8 instead, attacking my bishop on c7. I then took that bishop on b8, and this was the worst moment of the game.
After my capture Black had mate in six, starting with the rook taking my knight on b1 with check. My opponent took back on b8 with the queen instead, and the position was dead equal. We traded queens, Black took back on b8 with a rook on move 21 instead of checking on e2 (another blunder), I castled when the engine wanted the knight back on d2, and the game turned into a long rook endgame with me a pawn up.
Moment 3: the endgame, swinging both ways
The endgame was a sequence of chances for both of us. On move 40 I pushed f4 when taking on e5 with check was winning (+4.21 → +0.63). Black gave it back on move 41. On move 45 I made the clearest mistake of the ending.
Taking the pawn on g5 dropped the evaluation from +4.45 to +0.35. The engine wanted the a-pawn running. Black gave the chance back at once, taking my d4 pawn with the rook, and the same a-pawn went a6, a7 and a8. Black gave up the rook for it, I took back, and on move 49 it was my king and rook against Black's king and one pawn on d5.
Moment 4: the rook vs pawn endgame
Five moves later Black's pawn was on d4, Black's king on e2 next to it, my king on f4 and my rook on a1. The engine still scored it as winning for me. I walked my king to e4 to attack the pawn.
The tool describes my move as attacking an undefended pawn, and that was the whole idea in my head. But Black didn't need to defend it: the pawn went to d3, where its own king on e2 protected it. When I checked the position again for this post, the engine rated the position after my king move "about equal". It went d3, then d2, Black's king slipped to e1, and the pawn queened with check on move 58. My rook took the new queen, Black's king took my rook, and that was the draw.
The engine's move, rook to a2 with check, was the opposite order: push Black's king away from the pawn first, and bring my own king in after. When I look at it now, the same mistake runs through the whole game. At move 17, move 20 and move 55 I chose the move that attacked something over the move that kept control.
What the engine says about this game
My accuracy by phase was 93.2 in the opening, 67.2 in the middlegame and 78.2 in the endgame. Six blunders for me, seven for Black. The game analysis put four of my mistakes and blunders in the middlegame, four in the endgame, and none in the opening. Clock was never the excuse: at the decisive move 55 I had 9 minutes 40 seconds left, and I finished with more than eight minutes.
The lesson I took from game #284
The London System opening part is simple: when the opponent's bishop goes to b6 early and my queen is on b3, the a-pawn push can trap it or win the b7 pawn. That worked.
The bigger lesson is about converting. A rook vs pawn endgame is not automatically won, and the order of moves matters more than which piece "attacks". I now ask, before moving my king toward a pawn, whether the pawn can simply advance with its own king beside it. If it can, the rook checks or cuts the king off first.
What came next
I lost my next two games, the first against Raviraj13 (1072), and finished the week at 1023. Four weeks later, in game #350, I met hosseinkameli76 again in a London System, rated 1019 this time, and lost. The week after this game is week 10, and every week and milestone is on the journey page. You can replay game #284 on Chess.com.
If you want your own endgames checked this way, the moment found and the reason 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 Ke4 here, I wrote about Chess.com's game review not explaining moves.
Frequently asked questions
Can a rook vs pawn endgame be a draw?
Yes, and I drew one. In game #284 I had a rook against a lone pawn, and the engine called it winning for me until move 55. My king walked to the pawn instead of my rook checking first, the pawn advanced with its king beside it, and the game ended king against king.
What was the opening lesson from this London System game?
The opening went well: Black's bishop got stuck on the queenside and my a-pawn and queen won material by move 11. The lesson came later: two greedy captures (a pawn on c7 and the bishop on b8) gave nearly all of it back.
How many mistakes did I make in game #284?
By the engine's count: six blunders, two mistakes and four inaccuracies. My opponent made seven blunders and four inaccuracies. My accuracy by phase was 93.2 in the opening, 67.2 in the middlegame and 78.2 in the endgame.