Analysing My Recent Losses
I’ve been losing a lot lately and I wanted to understand why. Instead of reviewing each game individually, I ran all 7 of my recent losses through Stockfish at depth 18 and looked for patterns. The results are pretty revealing.
The Numbers
Seven games, a mix of rapid, blitz, and bullet. Here’s the damage:
| # | Opponent | Time Control | Blunders | Inaccuracies & Mistakes |
|---|---|---|---|---|
| 1 | 737knight | 10 min | 3 | 8 |
| 2 | VanDerTuesel | 2+1 Bullet | 1 | 5 |
| 3 | ramaa-2 | 2+1 Bullet | 1 | 2 |
| 4 | heythere111222 | 2+1 Bullet | 2 | 11 |
| 5 | Jah-Rastafarii | 10 min | 1 | 2 |
| 6 | Valet22600 | 10 min | 5 | 6 |
| 7 | ali1872637 | 3+2 Blitz | 1 | 8 |
14 blunders across 7 games, plus plenty of smaller inaccuracies and mistakes. The real question is: what kind of errors am I making?
Pattern 1: Not Seeing Available Captures
This is the biggest one. In 30% of my blunders, the engine’s best move was a simple capture that I didn’t play. In some cases, the same capture was available for multiple consecutive moves and I never played it.
Game 3 vs ramaa-2: dxe5 available for 3 moves straight
Black played 7…Nxe5 and I just… didn’t take back. My d-pawn could have captured on e5 for free. Instead I played Be4, putting my bishop on a square where it’s attacked by both the d5 pawn and the f6 knight. Black took it immediately with dxe4. Then after trading knights, I played Qxg4 taking a pawn, but the knight recaptured and I lost my queen too. Bishop and queen gone in three moves. I lost in 10.
Game 4 vs heythere111222: Bxg7 available for 4 moves straight
My position was already losing out of the opening. 8. d5 was bad, pushing a pawn into a position where it gets captured immediately. Then 9. c4 left my bishop on d3 hanging after the knight recaptured on d5 and hopped to b4. The trade Bxg7 Kxg7 was the engine’s top choice on moves 9, 10, 11, and 12, but I never took it. By move 15 I played h4, which allowed Black to push e4, forking my bishop on d3 and knight on f3. The position was already close to lost, but that fork sealed it.
Game 1 vs 737knight: Bxh8 in the endgame
This one hurts. I was in a drawn position and my opponent’s rook on h8 was completely hanging. My bishop on d4 could have just taken it. Instead I played Bg7, landing one square short of the rook. Almost certainly a mouse slip: the bishop went along the right diagonal but stopped on g7 instead of h8. One square away from winning a whole rook. Two moves later I got a second chance: after 38. Rb7+, my bishop on g2 could have taken the rook with Bxb7. I played Ke6 instead. Two chances to win a rook with a bishop capture, missed both.
Pattern 2: Queen Safety
15% of my blunders were queen moves. The worst:
Game 5 vs Jah-Rastafarii: Queen walks into a knight
The position was roughly equal. I played 12. Qd4 and Black immediately took it with Nxd4. Game over. The engine wanted Qe2, developing the queen to a safe square that also supports the centre. Instead I put my queen on the one square where it could be captured.
Game 7 vs ali1872637: Queen shuffling while a favourable trade sits on the board
My knight on e5 was attacking Black’s bishop on d7. The engine says Nxd7 on moves 17 and 18. It’s not winning material outright since Black recaptures with the queen, but the full line is Nxd7 Qxd7 Rxe8+ Rxe8: it trades down and removes Black’s active rook with a check, simplifying into a much more manageable position. Instead I moved my queen from f3 to g3, then to h3, then back to g3. Three queen moves while a favourable exchange sat on the board.
Game 7 vs ali1872637: Wrong capture
On move 24, my knight on e5 was under attack from the queen, pawn, and rook. It was going to die regardless. I played Bxd4 trying to defend it, but that was the wrong instinct. The engine says Qxf5: accept the knight is lost and grab Black’s bishop as compensation. After the follow-up trades it would be a rook and bishop each with Black up two pawns. Still losing, but the game goes on. Instead, Bxd4 took a pawn while my knight got captured for nothing, and the position fell apart.
Pattern 3: Ignoring Opponent Threats
Game 6 vs Valet22600: Failing to convert a winning position
I was up a piece in this game and needed to keep my king safe to convert the win. Instead I made it worse.
The engine says Nf3 here, developing the knight to a square where it defends and keeps the kingside solid. Instead I played g3, weakening the pawn structure in front of my king. After g3, Black’s best move is Rxg3+!, a rook sacrifice that rips open the kingside: fxg3 Qxg3+ Kh1 Qxh3+ and the attack is devastating. Black missed this and played Qe5 instead.
Even after Black missed the rook sacrifice, the engine still says Nf3. Instead I played Rac1, a move that does nothing to address the weak kingside. That took me from +4.2 to mate in 7. Black didn’t play accurately either, and the eval drifted back to +0.9 in my favour. But then I played g4, allowing Rxg4+ which forced me to exchange my queen for a rook. From there it was over.
Pattern 4: Missed Tactics
Game 2 vs VanDerTuesel: Missed Nxd4+ capture with check
This was a bullet game and I was in a time scramble: 10 seconds on my clock versus my opponent’s minute. I was winning (-5.0 eval) but the clock was the real enemy. On move 15, White’s knight on b5 was hanging and I could have taken it, but I played Rac8 instead. Three moves later, my knight on e6 could have played Nxd4+, capturing the bishop and giving check to the king on c2. The c3 pawn can’t recapture because it’s pinned to the king by my rook on c8. After Kd1, the knight hops to b3 attacking the rook on a1. White has to move the rook, and I’ve won a bishop for nothing. The full line is Nxd4+ Kd1 Nb3 Ra3 Ra8. Instead I played Rc4, which did nothing useful. The next move I missed Bxa4+, winning a pawn with check. The point of these missed tactics is the same: I’m not scanning for checks and captures before I move, and I’m not noticing when pins create opportunities.
The Verdict
My suspicion was right: these are mostly simple blunders. Looking at the data:
- 72% of blunders happened in playable positions. These aren’t positions where I was already lost. These are positions where the game was still competitive and I threw it away.
- 30% of blunders had a simple capture as the best move. I’m not scanning the board for captures before I move.
- The middlegame is my weakest phase (50% of blunders). My openings are okay, but once the position gets complicated I start missing things.
What I’m Going to Work On
One thing. Just one thing, because if I fix this it addresses most of these patterns:
Before every move, check all captures.
Can I take something? Can they take something? If the answer to either question is yes, I need to evaluate whether that changes my plan. Games 1, 2, 3, 4, and 7 all featured positions where I had a winning capture and didn’t see it. That’s 5 out of 7 games where checking for captures would have changed the result.
I need to always be asking myself: “What is my opponent threatening?” Games 5 and 6 were lost because I ignored threats. In Game 5 the knight was attacking my queen. In Game 6 the queen was threatening checkmate. Both times I played a move that had nothing to do with the threat.