Week 78: how to improve at chess tactics, starting with three missed knight forks
Week 78 was three games and one pattern. Between July 24 and 27, 2025, I won 1 and lost 2, and my rating barely moved: 1242 → 1241. But the engine found the same missed idea three times in two games: a knight fork, one knight jump attacking two or three pieces at once. If you're wondering how to improve at chess tactics, this week is my honest answer: the tactics I need to study are the ones I keep missing in my own games.
The week in numbers
| Games | 3 (2 Black, 1 White) |
| W / L / D | 1 / 2 / 0 |
| Rating | 1242 → 1241 (−1); 1257 after the win |
| Average accuracy | 73.6% (my opponents: 77.1%) |
| Blunders | 10 (my opponents: 11) |
| Openings | Modern Defense with 1.d4, Zukertort Variation, Modern Defense with 1.e4 (1 each) |
The week started well. In game 1620 against Etherion88 (1247), White pushed 3.f5 early, I took the pawn with my bishop, and I went on to win by checkmate on move 25 with 95.3% accuracy. The next two games were the forks.
Moment 1: the fork that hit three pieces
Game 1621 against archenemies500 (1263), White, a London setup (game). Black had just castled queenside, and the engine said that was a blunder: +5.70 for me. My knight was already on g5.
Two pawns were within reach of my knight, f7 and h7. Nxf7 is the fork: from f7 the knight attacks Black's queen on d6, the rook on d8 and the rook on h8, and Black can't save all three. I took the other pawn, on h7, and the evaluation swung from +5.70 to −2.82. Black then missed his best reply too, and the game went on for another 23 moves before I abandoned it. My best winning chance in that game had been 89.6%.
Moment 2: the same fork, offered twice
Game 1622 against AshokNanda_07 (1247), Black, Modern Defense (game). White had brought his queen out early to f3, with the bishop on c4, and his sixth move, Nge2, was a blunder: −6.76 for me.
Ne5 attacks the queen and the undefended bishop at the same time. White has to move the queen, and then the knight takes the bishop. I castled. White played 7.a3, which left the fork on the board, and the engine had me at −6.64 again. I played 7...e6 and missed it a second time. The game went on with mistakes on both sides, and I was checkmated on move 30.
Why I keep missing forks
Looking at both positions, the reason is clear. In 1621 I was looking at pawns: f7 and h7 both looked like "free pawn" captures, and I picked one without asking what the knight would attack from its new square. In 1622 I was following my opening routine, castle, develop, and didn't look at what the opponent's last move had left loose. A fork is easy to see once you ask the right question, and very easy to miss if you never ask it.
The lesson of the week
A tactics habit I can actually use in a 10-minute game: whenever the opponent's queen, king or an undefended piece stands a knight's jump from a square my knight can reach, check that square. In 1621 the check would have been "what does my knight attack from f7?" In 1622 it would have been "what does my knight attack from e5?" Both answers were "two or three pieces".
For training, these positions are worth more to me than random puzzles, because they're the patterns that appear in my own openings. That's how to improve at chess tactics in practice: collect your own misses and replay them until the pattern jumps out.
All games this week
| Game # | Date | Colour | Opponent (rating) | Result | Opening | Accuracy | Link |
|---|---|---|---|---|---|---|---|
| 1620 | Jul 24 | Black | Etherion88 (1247) | Win | Modern Defense with 1 d4 | 95.3% | game |
| 1621 | Jul 25 | White | archenemies500 (1263) | Loss | Queen's Pawn Opening Zukertort Variation | 63.7% | game |
| 1622 | Jul 27 | Black | AshokNanda_07 (1247) | Loss | Modern Defense with 1 e4 | 61.9% | game |
Previous: week 77 · Next: week 79 · All weeks · Turn the forks you missed into practice with Chess Coach's AI analysis.
Frequently asked questions
How do you improve at chess tactics?
Start with your own misses. In week 78 I missed the same pattern three times in two games: a knight jump that attacked two or more pieces at once. Positions from your own games are the patterns you actually face.
Why are knight forks so easy to miss?
Knights move in an L and don't attack along lines, so their targets don't look connected. In game 1622 my knight on c6 could jump to e5 and hit White's queen on f3 and bishop on c4, and I played other moves two turns in a row.
Do players at 1250 miss tactics too?
Constantly. In game 1622 White left the same fork available for two moves, and in game 1621 Black castled straight into a knight fork. Each player made at least four blunders in each of those games.