Week 38: how to get better at chess calculation, from captures to stalemates
Week 38 was a good week on paper: from September 23 to 29, 2024, I played 38 rapid games, won 20, lost 14, drew 4, and went from 999 to 1044, touching 1076 on the way. But the four draws include two stalemates in games where I had mate in two, and the three key mistakes the engine picked are all captures I didn't calculate. If you are trying to work out how to get better at chess calculation, this week is a list of exactly the calculations I skipped.
The week in numbers
| Stat | This week |
|---|---|
| Rating | 999 → 1044 (+45) |
| Games | 38 (#908–#945) |
| Wins / losses / draws | 20 / 14 / 4 |
| Blunders (my moves) | 80 (2.1 per game) |
| Average accuracy | 78.6 % |
| Most played openings | Queens Pawn Opening Zukertort Chigorin Variation (8), Queens Pawn Opening Accelerated London System (3), Giuoco Piano Game Giuoco Pianissimo Variation (3) |
The week's engine was my White repertoire. In the eight games Chess.com labelled "Zukertort Chigorin" (mostly 1.d4 d5 2.Bf4 Nc6 3.Nf3, plus similar move orders) I won seven and lost one, and overall I scored 12–4–3 with White. With Black it was 8–10–1. Only four of my 14 losses were recorded as abandoned this time; six ended by resignation and four by checkmate.
The ending was sour. After beating hastilychess (1077) in game 941 I reached 1076, then lost my last four games of the week and finished at 1044.
How to get better at chess calculation: three captures I didn't finish
Calculation, at my level, mostly means one thing: after my move, what is my opponent's best reply? Each of these moments is a capture where I stopped one move too early.
Game 908: giving a bishop for a pawn and a check
With White against wadk101 (1001), in a London System, the position was about equal (+0.39). Black's pawn on b5 was defended by the pawn on a6, but my bishop could take it with check.
Position.py's warning is in its first line: "but the bishop can now be captured". After 8...axb5 9.Qxb5+ I had two pawns for the bishop, and the engine's verdict was "black is winning" (-4.55). I had calculated the check, not the material. The engine's alternatives (Ne5, Nh4, a4) all kept "about equal". I lost the game.
Game 909: a pawn fork while my queen was hanging
With Black against AmbmToni (973), a Scotch Gambit, I was "completely winning" (-7.11). White's queen on c5 and mine on e7 were facing each other on the same diagonal, and I had a pawn on c3.
cxb2 is a lovely-looking pawn fork: position.py lists a double attack on the rook in the corner and the bishop. But my queen was still attacked, and White took it first: 11.Qxe7+ Ngxe7 12.Bxb2. The evaluation went from -7.11 to +0.53. Taking White's queen first, Qxc5, simply wins it: nothing on White's side can take back on c5, and the engine's line goes on 11.Nxc3, "black is completely winning". I still won the game by checkmate, 90 half-moves in.
Game 910: two pawn captures, one with an attack on the queen
With White against Mittaaaa (994), I was "completely winning" (+6.85) and had two pawns to take with my bishop.
Both are pawn captures, but only one comes with a threat. Bxc7 attacks the queen on d8, and in the engine's line 9...Qc8 10.Bxa6 I win the a6 pawn anyway (+7.55). After Bxa6 Black got time for Bd6 and the score dropped to +3.08, still "white is winning". I won the game by resignation, but the lesson is the same: compare captures by what they threaten, not just by what they take.
Two stalemates with mate in two
The draws hurt more than the losses. In game 912, with Black against Anish_Baraili (1031), I had promoted to a second queen. The engine shows mate in two on the board before my last move, and I played a queen move that left White's king with no legal moves: stalemate. I had 34 seconds on my clock, and 17 of my moves in that game were played with a low clock.
In game 931, with White against rafaci (1041), it happened again: mate in two according to the engine, and my pawn capture left Black's king without a single legal move. This time I had more than two and a half minutes left, so I can't blame the clock.
(The fourth draw, game 939, went the other way: I had only my king left against a queen, my opponent's clock ran out, and the game was drawn because I had no material to mate with.)
The lesson of the week
Every mistake on this page is a calculation I stopped one move too early. For captures, that means asking what does my opponent capture next? (games 908 and 909) and which capture also creates a threat? (game 910). For winning endgames, it means one extra question before every move: after this, does my opponent have a legal move? If the answer is no and they are not in check, it's stalemate. Two half-points from completely won positions would have been worth more rating than any opening I studied that month.
All games this week
| Game # | Date | Colour | Opponent (rating) | Result | Opening | Accuracy | Link |
|---|---|---|---|---|---|---|---|
| 908 | Sep 23 | White | wadk101 (1001) | Loss | London System | 76.8 % | game |
| 909 | Sep 23 | Black | AmbmToni (973) | Win | Scotch Game Scotch Gambit Haxo Gambit | 71.7 % | game |
| 910 | Sep 24 | White | Mittaaaa (994) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 91.4 % | game |
| 911 | Sep 24 | White | Arjunwho (976) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 87.7 % | game |
| 912 | Sep 24 | Black | Anish_Baraili (1031) | Draw | Queens Pawn Opening Blackmar Gambit | 90.3 % | game |
| 913 | Sep 25 | Black | dylanmalcolm (993) | Win | Ruy Lopez Opening Cozio Defense | 90.9 % | game |
| 914 | Sep 25 | Black | adhiwibowoo (1015) | Loss | Queens Gambit Declined Baltic Pseudo Slav Defense | 68.3 % | game |
| 915 | Sep 25 | White | GL0B3X (1004) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 97.1 % | game |
| 916 | Sep 25 | White | Yosward2024 (1036) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 93.9 % | game |
| 917 | Sep 26 | Black | LAUKIKAT (1038) | Loss | Bishops Opening Boi Variation | 58.1 % | game |
| 918 | Sep 26 | Black | MostefaiSaidRabah (1022) | Win | Giuoco Piano Game | 85.8 % | game |
| 919 | Sep 26 | White | bogdan86qwerty (1008) | Win | Queens Pawn Opening Horwitz Defense | 91.7 % | game |
| 920 | Sep 26 | White | Joenuts23234 (1024) | Win | Queens Pawn Opening Accelerated London System | 98.2 % | game |
| 921 | Sep 26 | Black | tournacc (1041) | Loss | Slav Defense Modern Quiet Schallopp Defense | 87.1 % | game |
| 922 | Sep 26 | Black | GuilhermeSL33 (1032) | Win | Giuoco Piano Game Main Line | 61.9 % | game |
| 923 | Sep 26 | White | treekiller4125 (1053) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 95.8 % | game |
| 924 | Sep 26 | Black | atomhuszti27 (1055) | Win | Bishops Opening Boi Variation | 79.3 % | game |
| 925 | Sep 27 | Black | MS0094 (1079) | Loss | Giuoco Piano Game Giuoco Pianissimo Variation | 55.7 % | game |
| 926 | Sep 27 | White | Mikhailpobeda (1045) | Loss | Queens Pawn Opening Accelerated London System | 97.7 % | game |
| 927 | Sep 27 | White | Rove85 (1063) | Loss | Queens Pawn Opening Zukertort Chigorin Variation | 81.6 % | game |
| 928 | Sep 27 | White | uovercast (1049) | Draw | Indian Game | 91.5 % | game |
| 929 | Sep 27 | Black | Abdul_Rauf_2005 (1013) | Win | Giuoco Piano Game Giuoco Pianissimo Variation | 60.8 % | game |
| 930 | Sep 27 | Black | Turash10 (1033) | Loss | Giuoco Piano Game | 69.9 % | game |
| 931 | Sep 27 | White | rafaci (1041) | Draw | English Defense | 75.1 % | game |
| 932 | Sep 28 | Black | shanoonmaldives (1044) | Loss | Giuoco Piano Game Giuoco Pianissimo Variation | 53.4 % | game |
| 933 | Sep 28 | White | Iamclaric (1030) | Win | English Defense | 97.6 % | game |
| 934 | Sep 28 | Black | BojanMAT (1039) | Win | Queens Pawn Opening Chigorin Variation | 66.4 % | game |
| 935 | Sep 28 | White | pep_jai (1017) | Win | Indian Game | 66.4 % | game |
| 936 | Sep 28 | Black | sobe_2009 (1037) | Win | Ruy Lopez Opening Morphy Defense Cozio Defense | 87.0 % | game |
| 937 | Sep 28 | White | Raza777kk (1029) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 80.5 % | game |
| 938 | Sep 28 | Black | jannazai (1069) | Loss | Scotch Game | 75.6 % | game |
| 939 | Sep 28 | White | dadmuld (1053) | Draw | Queens Pawn Opening Krause Variation | 81.0 % | game |
| 940 | Sep 29 | White | atsiro (1069) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 84.3 % | game |
| 941 | Sep 29 | White | hastilychess (1077) | Win | Queens Pawn Opening Horwitz Defense | 90.7 % | game |
| 942 | Sep 29 | Black | WojciechPaszkowski (1086) | Loss | Giuoco Piano Game Evans Accepted McDonnell Defense | 56.5 % | game |
| 943 | Sep 29 | Black | consultoramakur (1085) | Loss | Scotch Game | 55.3 % | game |
| 944 | Sep 29 | White | Ahireyogesh (1057) | Loss | Queens Pawn Opening Accelerated London System | 52.2 % | game |
| 945 | Sep 29 | Black | Sanaa990 (1071) | Loss | Queens Pawn Opening | 80.2 % | game |
Before this: week 37 · After this: week 39 · The journey from 631 to 1385 · Review your own games like this
Frequently asked questions
How do I get better at chess calculation as a beginner?
My week 38 games suggest starting small: for every capture, calculate the opponent's best reply, especially captures and checks. Two of my three key mistakes were captures that ignored a capture coming back.
How do you stalemate by accident?
By making a move that leaves the opponent with no legal moves while they are not in check. In game 912 I had two queens and mate in two, and in game 931 I had mate in two again; both times I played a move that ended the game in a draw.
Did my rating still go up this week?
Yes, from 999 to 1044, with a peak of 1076 after game 941. With White I scored 12 wins, 4 losses and 3 draws.