How to get better at chess calculation: the mate in 3 I missed in week 25

In week 25 (June 25–30, 2024) I played 22 rapid games, all at 10 minutes, and finished with 11 wins, 9 losses and 2 draws, up from 1020 to 1038. It included a six-game winning streak on June 26. But the game I keep thinking about is the first one of the week, where Stockfish found a mate in 3 that I didn't see, and I lost. If you've wondered how to get better at chess calculation, that one position is the best case I have for where to start.

The week in numbers

For once, Black was my better colour: 6 wins, 3 losses and a draw, against 5–6–1 with White. Wednesday, June 26 was the good day, twelve games with 8 wins, 3 losses and a draw, including the six-game streak from #608 to #613. My London setups were steady (London System 2–1–1, Accelerated London 2–0).

The rest was familiar. In seven games I had at least an 80 % chance of winning and didn't win. Seven of my nine losses are logged as abandoned games. Game #614 against rudrabidave was a 66-move loss on time with 4 seconds left on my clock, from a position where my winning chance had peaked at 93.9 %. And the second draw of the week was a stalemate, in game #627 against ECaris (1046). This time I was the one defending a king and pawn ending, so stalemate saved me half a point, but only after I had let a position with a 95.4 % winning chance slip earlier in the game.

Three forcing moves

Mate in 3, and I played a rook move

Game #607, White against madhusudhanbachalakuri (1046). Move 34, and Stockfish says White has a forced mate in three.

. . . . . . r k
. . . . . . p p
. . . . . . . .
. . . . . Q P .
. . p B . P . .
. . P . P . K P
r q . . . . . .
. . . R . R . .

The mate goes 34.Bxg7+ Rxg7 35.Rd8+ Rg8 36.Qe5#, with the queen delivering mate on the long diagonal. Every move is a check. The first one gives away my bishop, and I think that's exactly why I never looked at it. Instead I played 34.Rf2, which attacks Black's queen on b2 with the rook on a2 behind it, a double attack. But Black has an answer: 34...Qxf2+, taking my rook with check. My king can't take back, because the queen on f2 is protected by the rook on a2 along the second rank. A forced mate became 0.00, about equal, and I abandoned the game three moves later (game on Chess.com).

A capture that attacks two pieces

Game #608, Black against Michelpaban (1001), in an East Indian Defense. About equal (+0.55), and White's e4 pawn was within reach of my knight.

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

13...b5 supports my c4 pawn, but the b5 pawn itself can be taken. Stockfish preferred the capture 13...Nxe4. The knight takes a pawn, attacks White's knight on c3 and bishop on g5, and by leaving f6 it opens the diagonal for my bishop on g7, which now also hits c3. One move, several threats. +0.55 became +3.73 after b5, while 13...Nxe4 kept White's edge small. I won this game anyway, by checkmate (game on Chess.com).

A check that let Black block with tempo

Game #609, White against Hilal_Ahmed (1028), in my Zukertort Chigorin setup. I was winning, +5.66, with my queen on b7 in Black's camp.

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

9.Qb5+ looks like a perfect double attack: check, and the knight on a5 is attacked too. But Stockfish's answer for Black is to block with 9...c6, and that pawn attacks my queen. Then I have to move the queen again, and the knight is no longer simply lost. The evaluation fell from +5.66 to +2.12. 9.Qa6 attacks the same knight without giving Black a useful block. My opponent blocked with 9...Nc6 instead, a blunder that Stockfish scores at +9.73 for me. I won this one by checkmate after 70 moves (game on Chess.com).

The lesson of the week

All three moments are about forcing moves: checks, captures and threats. I missed a sacrificial check, I missed a capture with multiple threats, and I played a check without asking how Black could answer it. So the method I'm taking from this week, as my own answer to how to get better at chess calculation, is an order of search: every check first, then every capture, then every threat, for me and for my opponent, and for each one the opponent's replies, including blocks that hit back. A sacrifice that "loses" a piece still goes on the list. In game #607 the whole mate was three checks long, and I'd have found it by looking at checks before anything else.

All games this week

Previous: week 24 · Next: week 26 · Overview: the journey hub · Find your own missed mates: review your own games like this.