Week 107: stuck at 1200 elo again, three games and three different mistakes
Three games in three days, February 23 to 25, 2026, and three losses. My rating went from 1228 to 1204 (−24), the lowest it had been since October 27, 2025, and 181 points below my peak of 1385. I think "stuck at 1200 elo" describes it exactly: four weeks after October 27 I had climbed from 1204 to my peak, and now I was back where that climb started. The engine review is useful precisely because it doesn't care about that. It shows three games and three separate mistakes, each one small enough to fix.
The week in numbers
| Games | 3 |
| Wins / losses / draws | 0 / 3 / 0 |
| Rating | 1228 → 1204 (−24) |
| Average accuracy | 75.2 % |
| Blunders | 11 |
| Openings | Modern Defense with 1 e4 (1), Queen's Pawn Zukertort Chigorin (1), Pirc Geller System (1) |
Accuracy of 71.6 %, 74.8 % and 79.1 % — not disasters. Eight of the eleven blunders came in the middlegame. Two games ended in checkmate against me, and in all three the engine had me better at some point: −1.25 in the first, +2.59 in the second and −6.08, completely winning, in the third.
Mistake one: a knight too far forward (game 1986)
Monday, Black against Shtondik0007 (1225) in a Modern Defense. White had pushed a pawn to f5, attacking my pawn on g6.
12...Nd4 puts the knight on a beautiful square, but the engine shows White's reply 13.Nd5 and then 14.f6, a pawn that attacks my bishop on g7 (+2.18). With 12...e6 I challenge the f5 pawn directly and the game is only slightly better for White (+0.53). Four moves later I made it worse: 16...Bxb2 took a pawn with my dark-squared bishop and attacked the rook on a1, but the engine rated it +6.44, "white is winning". It cost me two moves (17.Rb1 Bg7, back where the bishop started), and White used them: 18.Qh5, 19.Ng5 and 20.Qxh7 mate (game).
Mistake two: recapturing with the wrong pawn (game 1987)
Tuesday, White against ajayanaji (1213). On move 19 I could take Black's pawn on b4 with either the a-pawn or the c-pawn.
Both moves are described the same way: "pawn takes the pawn (b4)". The engine's numbers are not the same at all: 19.cxb4 +2.07 (clearly better for White), 19.axb4 +0.44 (about equal). After axb4 my pawn on c3 stands alone, and the engine's first reply, 19...Ne4, attacks it — it has no defender. After cxb4 my pawn on a3 protects the one on b4. I stayed better for a long time anyway (+2.49 on move 35), then two knight moves, 35.Ne7+ and 36.Nc6, turned it around, and Black mated me on move 40 (game).
Mistake three: not cashing in (game 1988)
Wednesday, Black against Abomalek23 (1217) in a Pirc. My knight had already taken a pawn on g2 earlier and I was completely winning (−6.08).
23...Qxc2 takes a pawn, attacks White's queen on d1 and the pawn on b2, and the engine's line trades queens into an endgame where I stay winning (−5.91). I played 23...d5 instead, and the engine's first answer is 24.Bxb4, White's bishop taking my b4 pawn (−1.72). The game slowly turned, and two blunders on moves 38 and 39 (+7.34 after 39...c6) finished it; it was abandoned on move 40 (game).
The lesson of the week
Being stuck at 1200 elo, at least for me, isn't about knowing too little. All three mistakes are ideas I know: keep the fianchetto bishop near the king, recapture towards the pawns that can protect each other, and when completely winning, trade queens and take the free pawn. What I lack is the habit of stopping for ten seconds at the moment the game is good for me.
All games this week
| Game # | Date | Colour | Opponent (rating) | Result | Opening | Accuracy | Link |
|---|---|---|---|---|---|---|---|
| 1986 | 2026-02-23 | Black | Shtondik0007 (1225) | Loss | Modern Defense with 1 e4 | 71.6 | game |
| 1987 | 2026-02-24 | White | ajayanaji (1213) | Loss | Queen's Pawn, Zukertort Chigorin | 74.8 | game |
| 1988 | 2026-02-25 | Black | Abomalek23 (1217) | Loss | Pirc, Geller System | 79.1 | game |
Previous: week 106 · Next: week 108 · The journey so far · Get your own games reviewed like this
Frequently asked questions
Why am I stuck at 1200 elo?
In my case the losses weren't from bad openings. In all three games this week I had a good position at some point (−1.25, +2.59 and −6.08 in the engine's numbers) and gave it back with a single move.
How far did my rating drop from the peak?
From 1385 on November 23, 2025 to 1204 on February 25, 2026: 181 points. The last time I had been at 1204 was October 27, 2025.
Which pawn should recapture?
The one that keeps your pawns together. In game 1987 both 19.axb4 and 19.cxb4 take on b4, but cxb4 kept me clearly better (+2.07) and axb4 left my c3 pawn alone (+0.44).