Week 104: London System mistakes I made six games in a row
Twelve games from January 19 to 24, 2026, and only 4 wins against 8 losses. My rating dropped from 1292 to 1261 (−31). With White I played the London System setup (d4 and Bf4 on move 2) in all six games and scored 2–4, so this is a page about London System mistakes — mine, with the engine's verdict next to each one. With Black it was no better: 2–4 as well.
The week in numbers
| Games | 12 |
| Wins / losses / draws | 4 / 8 / 0 |
| Rating | 1292 → 1261 (−31) |
| Average accuracy | 75.4 % |
| Blunders | 35 |
| Most played openings | London System (3), Pirc Classical (2), Modern Defense with 1 e4 (2), East Indian London (1), Queen's Pawn Zukertort (1) |
The rating touched 1300 after my first game on Monday and never came back. Five of the eight losses were games where the engine flags a winning position I let go. My wins averaged 86.8 % accuracy, my losses 69.7 %, and 20 of the 35 blunders were in the middlegame.
London System mistakes, one by one
Mistake 1: the knight that wandered off (game 1965)
Monday morning, White against siva2105 (1263), who set up with g6, Bg7, Nf6 and castling. My knight went to g5 on move 10, Black chased it with ...h6, and I had to decide where it went next.
On f3 the knight stands on the long diagonal from b7 to g2 and blocks it. On h3 it doesn't, and the engine's reply to 11.Nh3 is 11...Bxg2: the bishop takes my g2 pawn and attacks my rook in the corner, which has no defender (−4.00, "black is winning"). After 11.Ngf3 Black is only slightly better (−0.97). The earlier 10.Ng5 was already an error (−0.24 → −1.93). My opponent played 11...c4 and abandoned after 12.Qc2, so I got the point I didn't deserve (game).
Mistake 2: trading my best knight (game 1966)
An hour and a half later, against pawnpawn93 (1331), I had a good London position: my bishop on b5 pinned the knight on c6, my queen stood behind the bishop on a4, and my knight on e5 attacked the pinned knight as well.
9.dxc5 attacks the queen on b6 and keeps the pressure (+2.24, "clearly better for white"). 9.Nxc6 cashes in immediately: it's a trade, Black takes back, and my advantage is gone (−0.39 after 9...Nxc6). The game went on for 46 moves, and it hurts to write the end: I was +14.98 with a pawn on b7 about to become a queen, Black checked me with 45...Rc3+, and I answered 46.Kf4, walking into 46...Rf3 mate. 46.Kd2 attacked the rook and kept a completely winning position. I had 64 seconds left (game).
Mistake 3: grabbing the rook (game 1975)
On Friday, against ANANTGOLA (1373), I was completely winning on move 11 (+8.44) and took a rook in the corner with 11.Bxa8. The position.py check explains the problem in one line: the capture "stops defending the queen (a4), which can now be captured". My bishop on c6 had been standing between Black's bishop on d7 and my queen. 11...Bxa4 followed, then 12.Nbd2 was another blunder, and the game was abandoned on move 12. 11.Qc2, simply moving the queen first, kept me completely winning (+7.93) (game).
And with Black
The Pirc gave me no comfort. Against Joob07 (1309) I was clearly better (−2.37) when 14...c5 let it go; 14...Nxd3, taking White's bishop, kept the edge (−2.27). I lost that one too, and that move was only the first of my five blunders in the game. My two wins with Black were clean, though: 91.1 % against dr-Thursday (1248) with a checkmate, and 92.4 % against TaL-FromNepal (1254).
The lesson of the week
The London System is a setup, not a plan that plays itself. The mistakes that cost me points were never about the opening moves; they were about one piece moving and leaving something behind — the f3 knight leaving g2, the c6 bishop leaving my queen, my king leaving safe squares near the end. In the London, my pieces protect each other in a chain. Before I move one, I should ask what link I'm removing from that chain.
All games this week
| Game # | Date | Colour | Opponent (rating) | Result | Opening | Accuracy | Link |
|---|---|---|---|---|---|---|---|
| 1965 | 2026-01-19 | White | siva2105 (1263) | Win | Indian Game, East Indian London | 80.3 | game |
| 1966 | 2026-01-19 | White | pawnpawn93 (1331) | Loss | Queen's Pawn, Zukertort | 52.3 | game |
| 1967 | 2026-01-19 | Black | Joob07 (1309) | Loss | Pirc, Classical | 79.3 | game |
| 1968 | 2026-01-20 | White | magnusorsomething (1290) | Loss | London System | 52.3 | game |
| 1969 | 2026-01-21 | Black | dr-Thursday (1248) | Win | Modern Defense with 1 e4 | 91.1 | game |
| 1970 | 2026-01-21 | White | samir629 (1295) | Win | London System | 83.4 | game |
| 1971 | 2026-01-21 | White | alian9090 (1280) | Loss | London System | 86.3 | game |
| 1972 | 2026-01-22 | Black | AdityaTiwary28 (1291) | Loss | Modern Defense with 1 e4 | 72.0 | game |
| 1973 | 2026-01-22 | Black | TaL-FromNepal (1254) | Win | Old Benoni Defense | 92.4 | game |
| 1974 | 2026-01-23 | Black | nasira43 (1344) | Loss | Modern Defense with 1 d4 | 87.9 | game |
| 1975 | 2026-01-23 | White | ANANTGOLA (1373) | Loss | Queen's Pawn, Zukertort Chigorin | 54.5 | game |
| 1976 | 2026-01-24 | Black | beyaz1fil (1270) | Loss | Pirc, Classical | 72.8 | game |
Previous: week 103 · Next: week 105 · The journey from 631 to 1385 · Review your own games like this
Frequently asked questions
What are common London System mistakes?
In my six London games this week: moving the f3 knight away so the long diagonal to g2 opens, trading my best knight too early, and grabbing material with a piece that was shielding my queen.
Is the London System good for beginners?
It gives you a familiar setup, but it doesn't stop tactics. This week my London games scored 2 wins and 4 losses, and three of the losses came from positions the engine rated as winning for me.
Why did the f3 knight matter so much?
In game 1965, after 11.Nh3 the engine's reply is 11...Bxg2, taking my g2 pawn and attacking my rook (−4.00). Going back with 11.Ngf3 kept the game close (−0.97).