Week 40: hanging a knight in a winning endgame
Week 40, October 7–12, 2024: 20 rapid games, 8 wins, 11 losses and 1 draw, and a rating slide from 1042 to 1017. The three key moments are knight moves; two put the knight on a square where it could be captured. Hanging a knight in a winning endgame is the clearest example below.
Hanging a knight in a winning endgame: the week in numbers
| Games | 20 (#1006–#1025) |
| Wins / losses / draws | 8 / 11 / 1 |
| Rating | 1042 → 1017 (-25) |
| Average accuracy | 76.1 % |
| Blunders (my moves) | 53 (2.6 per game) |
| Most played openings | Queens Pawn Opening Zukertort Chigorin Variation (4), Kings Pawn Opening (3), Queens Pawn Opening Horwitz Defense (2) |
With White I went 6–4. With Black I won only 2 of 10 (7 losses and a draw), and every game Chess.com filed under "King's Pawn Opening", "Leonardis", "Center Game", "Bishop's Opening" or "Three Knights" ended in a loss for me. My best game came with Black anyway: game 1010 against gholamimsn49 (1018), a game Chess.com labels Slav Defense, with 98.0 % accuracy and not a single inaccuracy.
I made 53 blunders, 30 of them in the middlegame and 18 in the opening. Seven of my 11 losses were recorded as "game abandoned". In four of the losses, and in the draw, the engine had given me an 80 % or better chance of winning at some point.
Three knight moves and their exposed landing squares
A piece "hangs" when it stands where the opponent can take it for free, or for less than it is worth. Position.py, the tool I use to check every board in this diary, has a plain phrase for it: "but the knight can now be captured". Here are the three moves.
Game 1008: the knight that walked next to the king
The worst one first. With White against Bojan_Djordjevic (1063), I had reached an endgame of rook, knight and three pawns against rook and four pawns, and the engine said +4.76, "white is winning".
Nf7 attacks the rook. It also lands on a square next to the black king, and the king just took it: 38...Kxf7, and the evaluation went from +4.76 to -2.46. The engine's alternatives (b4, Kf2, h3) all kept "white is winning". I lost the game, and it went down as abandoned. A king is a piece that captures too, and in endgames it's often the most active one.
Game 1007: the fork square covered by a rook
With White against sika_mi16547 (1002), I was at +1.93 and saw what looked like a fork.
Position.py describes Nb4 as a double attack: the queen on d5 and the undefended pawn on a6. But the black rook on b8 looks straight down the open b-file, and the engine's reply is Rxb4: the rook takes the knight, I take the rook back with the c-pawn, and Black's queen takes my a2 pawn next to my king. The score goes to about equal (-0.03). My opponent played 24...Qa5 instead, and I won the game by resignation. The quiet b3 kept "clearly better for white".
Game 1006: chasing the knight instead of saving the bishop
With Black against NgocTuyen99 (1060), the position was about equal. White's knight on b5 was attacking my bishop on d6, and I attacked the knight back with a pawn.
Counter-attacking felt active, but the knight doesn't need to retreat: the engine's reply is Nxd6, taking my bishop. I can take back with the c-pawn, yet the engine rates the result "clearly better for white" (+2.25 in its line 10.Nxd6 cxd6 11.f4). Ne8 kept the game about equal. My opponent played 10.Nc3 and the moment passed, but I lost the game later.
The lesson of the week
None of these were deep tactics. In two of them a piece of mine landed on a square the opponent could take, and in the third I ignored that my own piece was already attacked. The habit I'm working on is a square check before I let go of a piece: which enemy pieces can reach that square, counting the king and any rook on an open file? And when my piece is attacked: is my counter-attack faster than the capture? In game 1008 that check takes two seconds and saves a won endgame. Learning how to stop hanging pieces in chess is less about talent and more about doing that check every single move.
All games this week
| Game # | Date | Colour | Opponent (rating) | Result | Opening | Accuracy | Link |
|---|---|---|---|---|---|---|---|
| 1006 | Oct 7 | Black | NgocTuyen99 (1060) | Loss | Kings Pawn Opening | 80.1 % | game |
| 1007 | Oct 7 | White | sika_mi16547 (1002) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 88.7 % | game |
| 1008 | Oct 7 | White | Bojan_Djordjevic (1063) | Loss | Queens Pawn Opening Zukertort Chigorin Variation | 79.0 % | game |
| 1009 | Oct 7 | White | mohamed3392 (1004) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 86.6 % | game |
| 1010 | Oct 8 | Black | gholamimsn49 (1018) | Win | Slav Defense Modern Quiet Schallopp Defense | 98.0 % | game |
| 1011 | Oct 8 | Black | funkoo19 (1059) | Loss | Kings Pawn Opening | 71.7 % | game |
| 1012 | Oct 8 | White | ARW1983 (1030) | Loss | Queens Pawn Opening Krause Variation | 49.9 % | game |
| 1013 | Oct 8 | White | cbaadfffddfftrrdswsddfghh (1060) | Loss | Queens Pawn Opening Zukertort Variation | 60.9 % | game |
| 1014 | Oct 8 | Black | 307pez (1027) | Loss | Center Game Accepted Normal Variation | 66.8 % | game |
| 1015 | Oct 9 | Black | oma425r (1040) | Draw | Giuoco Piano Game | 75.9 % | game |
| 1016 | Oct 9 | Black | Apshu3015 (1009) | Loss | Kings Pawn Opening | 73.0 % | game |
| 1017 | Oct 9 | White | rayan19990 (1008) | Win | Queens Pawn Opening Horwitz Defense | 78.8 % | game |
| 1018 | Oct 10 | White | yashchikhalkar (1024) | Win | Queens Pawn Opening Zukertort Chigorin Variation | 61.7 % | game |
| 1019 | Oct 10 | Black | nkism_19 (1021) | Loss | Kings Pawn Opening Leonardis Variation | 77.2 % | game |
| 1020 | Oct 10 | White | garda_boy (1039) | Loss | Indian Game | 75.7 % | game |
| 1021 | Oct 10 | Black | Cwells26 (1012) | Loss | Bishops Opening | 77.5 % | game |
| 1022 | Oct 11 | White | yuvraj-singh2008 (985) | Win | Queens Pawn Opening Accelerated London System | 94.4 % | game |
| 1023 | Oct 11 | White | eldichamp (1000) | Win | Queens Pawn Opening Horwitz Defense | 74.1 % | game |
| 1024 | Oct 12 | Black | lalitraut (1011) | Loss | Three Knights Opening | 68.5 % | game |
| 1025 | Oct 12 | Black | mvhMSL (1003) | Win | Kings Pawn Opening Tayler Opening | 84.4 % | game |
Previous week: week 39 · Next week: week 41 · Back to the journey · Review your own games like this
For a general method beyond this week's games, see the piece-safety routine.
Frequently asked questions
How did hanging my knight change a winning endgame?
In game 1008 my knight could be captured after I moved it next to the black king. The engine assessment changed from +4.76 to -2.46, and I lost the game.
What does hanging a piece mean?
It means leaving a piece where the opponent can take it for free or for less than it is worth. The engine's descriptions say it directly: 'the knight can now be captured'.
Can one hanging piece really lose a won endgame?
Yes. In game 1008 I was at +4.76, 'white is winning', and after my knight was taken the score was -2.46. I lost that game.