Week 39: sixty rapid games with no rating gain

Week 39 covers 60 rapid games from September 30 to October 6, 2024: 29 wins, 29 losses and 2 draws, with my rating moving from 1044 to 1042. I touched 1105 during the week and then gave the gain back. Sixty rapid games with no rating gain make a useful case for inspecting the moves behind the results.

Sixty rapid games with no rating gain: the week in numbers

Those 154 blunders are not spread evenly. In my 29 wins I blundered 39 times, about 1.3 per game; in my 29 losses, 108 times, about 3.7 per game. And 109 of the 154 came in the middlegame. Even more painful: in 19 of the 31 games I didn't win, the engine had given me an 80 % or better chance of winning at some point.

What sixty games showed about my repeated mistakes

I split the week by day to see whether volume itself hurt:

Thursday was the high point: 14 games, 9 wins, and a peak of 1105 after game 983. Friday undid most of it with six losses in eight games, and Sunday finished the job: four losses in a row at 69.2 % accuracy, three of them recorded as "game abandoned". In total 14 of my 29 losses were abandoned rather than resigned or played out.

The clock was not the issue. Only two of the 60 games ended on time, one win (game 951, with 12 seconds left) and one loss. Ten-minute games gave me enough time; I didn't use it to check my moves.

The openings were familiar. With White I played my 1.d4–Bf4 setup in its usual disguises: "Zukertort Chigorin" seven times (3 wins, 4 losses), "London System" and, six times, the Englund Gambit (1.d4 e5), where I scored 4 wins and 2 losses. With Black against 1.e4 I lost all four games labelled "King's Pawn Opening" and won all three labelled "Giuoco Piano".

Three winning positions that slipped

Game 946: a knight check instead of the queen raid

With Black against abdoelbazly79 (1052), the engine had me at -6.29, "black is completely winning". My queen sat on a8 and White's queenside knight on b1 had never moved.

q . . . . r k .
. . . . . p . p
. . . p . p b .
. p n . p . . .
. . . . P . . .
. . P P . . Q P
. . . . N P P .
. N . . K . . R

A check is always tempting. But position.py adds the warning: "the knight can now be captured", and White's queen took it, 19.Qxd3, "clearly better for white". Qa2 instead attacks the knight on b1, which had no defender, and the knight on e2; the engine's line wins the b1 knight a move later. The game ended in a draw by repetition.

Game 949: the king walk back to e8

With Black against joecanada83 (1016), White had sacrificed the bishop on f7 (1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ Kxf7 5.Nxe5+), my king was on f7, and the engine had me well ahead at -3.83. White's knight on e5 had just given check and was attacking my knight on c6.

r . b q . . n r
p p p p . k p p
. . n . . . . .
. . b . N . . .
. . . . P . . .
. . . . . . . .
P P P P . P P P
R N B Q K . . R

I moved my king back home. The engine simply takes the knight: after Nxe5 6.Qh5+ Kf8 7.Qxe5 d5 the engine still says "black is winning". After Ke8, White's best is Nxc6, a capture that also attacks my queen, and the score goes to about equal (+0.10). My opponent played 6.Qh5+ instead and I won the game by resignation, but that was their mistake, not my skill.

Game 950: moving the queen when the rook sacrifice won

With White against Prestocat (1054), in a London System, I was at +6.55. Black's pawn on b5 was attacking my queen on a4, and my rook on g1 faced the g7 pawn in front of Black's king.

r . . q r . k .
. n p . b p p p
p . . . p . . .
. p . p N . . .
Q P . P . B . .
. . P . P . . b
P . . . B P . P
. . K R . . R .

I saved my queen. Position.py describes Rxg7+ as a rook capture of "one of the pawns protecting the king" with check, a double attack that also hits f7 and the bishop behind it, and adds "the rook can be captured, but the line shows why it still works": 18.Rxg7+ Kxg7 19.Rg1+ Bg5 20.Bxg5, "white is winning". After Qc2 the engine's number was -1.06. My opponent didn't punish it, and I won the game by checkmate, with 7 blunders on my side along the way.

The lesson of the week

All three key moments are the same mistake: in a winning position, I answered the most obvious thing on the board — give a check, run with the king, save the queen — instead of looking for the move that keeps the initiative. Doing that 60 times in a week doesn't build the habit; it repeats the mistake 154 times. If I want to get better at rapid chess, the lesson is to play fewer games and check each one with the engine before the next. The data says I did play much less over the next three weeks: 20 games, then 8, then 9.

All games this week

Previous: week 38 · Next: week 40 · Journey hub · Review your own games like this

For a general method beyond this week's games, see the rapid improvement guide.