Week 63: three losses in the Pirc Defense, Modern Defense Geller System
Three weeks of climbing got me back over 1300; one week took it away again. Between March 24 and 30, 2025 I played 9 rapid games, won 3 and lost 6, and dropped from 1307 to 1282 (−25). I never got back to 1300 during the week: game 1405 took me to 1299 and that was the ceiling. With Black I went 1–5, and three of those losses came in the same opening, the one Chess.com labels the Pirc Defense, Modern Defense Geller System. Here's what happened in each.
The week in numbers
| Rapid games | 9 |
| Wins / losses / draws | 3 / 6 / 0 |
| Rating | 1307 → 1282 (−25) |
| Record with White / Black | 2–1 / 1–5 |
| Average accuracy | 70.7% |
| Blunders (mine) | 31 |
| Most played openings | Pirc Defense Modern Defense Geller System (3), Queen's Pawn Opening Zukertort Variation (2), Queen's Pawn Opening Accelerated London System (1) |
This was one of the roughest weeks of the winter by accuracy: 70.7% on average, with 31 blunders in 9 games. None came under one minute; every one of them had time behind it. On paper the opponents were about my level (average 1291); five of them had a higher rating than mine and I lost all five, and the three wins, all by checkmate, came against lower-rated players. With White I went 2–1.
The Geller System, three games
All three started with 1.e4 g6 2.Nf3, followed by my usual ...d6 or ...Nf6 and ...Bg7. Same setup as my Pirc and Modern games, different label.
- Game 1405 against DSweetScience (1301): loss. I was winning by more than five pawns twice and lost anyway (below).
- Game 1410 against inv45ion (1285): loss in 15 moves. 12...Bxd4 took a pawn with my dark-squared bishop and turned −0.23 into +4.30; the engine preferred to bring the bishop back to g7 with 12...Bg7.
- Game 1413 against Grandja (1310): loss. My queen went on a raid: 8...Qxb2, 9...Qxc3, 11...Qxa2, and 12...Qxa1 took a rook, but then 13.Qxa1 took my queen. The engine flags 11...Qxa2 as the blunder (−2.72 to +4.55); it wanted the queen back home with 11...Qd8.
I should have learned from week 58, where my queen went pawn-grabbing in game 1354 and I was mated. Apparently I hadn't.
Game 1405: the same capture I missed in week 60
With Black against DSweetScience (1301) (game on Chess.com), after nine moves I was already clearly better, −3.76.
White's dark-squared bishop had gone to e3 and the knight was still on b1, so nothing protected b2. The engine line is 9...Bxb2 10.Nbd2 d4 11.Bf4 Bxa1: my bishop takes the pawn, then the rook in the corner. I played the quiet 9...e6, and −3.76 became −0.48. This is exactly the capture I missed in game 1365, three weeks earlier: the same bishop on g7, the same pawn on b2, the same rook on a1.
It got worse. After 10.Nc3 the engine rated it −5.29 for me again, and wanted 10...d4, a pawn push that attacks both the knight on c3 and the bishop on e3. I played 10...Nf6 (−0.25). Then, on move 19, at −5.65, I played 19...Rd8 and the evaluation went to +14.93. My peak winning chance in this game was 92.6%. It ended as abandoned.
Game 1406: the open file I ignored
With White against wali_M (1295) in an Accelerated London (game link), the position was equal, +0.20. The a-file was open, with my rook on a1 facing Black's rook on a8.
I took the knight on c4 with my bishop from f1. The engine wanted Rxa8 first. After my Bxc4, the engine's best reply is 12...Rxa1+, the rook taking mine with check, and its line continues 13.Ke2 Rxh1: with my bishop gone from f1, Black's rook can also reach my other rook. +0.20 became −8.41. My opponent missed it and played 12...dxc4, but I lost this game later anyway, after 15.Nxc6 and four more blunders.
(The third key moment, 11...e6 in game 1407 against Punks1234 (1286), was a Modern Defense: it allowed 12.Nxd5, which also uncovers White's bishop on d2 against my queen on a5. The engine wanted 11...Qd8; the evaluation went from +0.18 to +2.20.)
The lesson of the week
The bishop on g7 again. For the second time in three weeks, the engine's best move was a capture on b2 with it, attacking an undefended rook on a1, and for the second time I played a quiet move. So this is my new rule for every g6 game, Geller, Pirc or Modern: whenever White's queenside pieces haven't moved, check whether my bishop on g7 can take on b2 and what it hits next. And its mirror image for my queen: if she is taking pawns on the queenside, I count how many moves it takes her to get back. In game 1413 the answer was "too many".
All games this week
| Game # | Date | Colour | Opponent (rating) | Result | Opening | Accuracy | Link |
|---|---|---|---|---|---|---|---|
| 1405 | Mar 24 | Black | DSweetScience (1301) | Loss (abandoned) | Pirc Defense Modern Defense Geller System | 49.9 | game |
| 1406 | Mar 25 | White | wali_M (1295) | Loss (abandoned) | Queen's Pawn Opening Accelerated London System | 70.0 | game |
| 1407 | Mar 25 | Black | Punks1234 (1286) | Loss (abandoned) | Modern Defense Standard Line | 73.5 | game |
| 1408 | Mar 26 | White | mostar14 (1257) | Win (checkmate) | Queen's Pawn Opening Zukertort Variation | 86.0 | game |
| 1409 | Mar 26 | Black | Azex290788 (1272) | Win (checkmate) | Modern Defense with 1 e4 | 79.2 | game |
| 1410 | Mar 27 | Black | inv45ion (1285) | Loss (abandoned) | Pirc Defense Modern Defense Geller System | 75.0 | game |
| 1411 | Mar 27 | White | givaborges23042001 (1287) | Win (checkmate) | Queen's Pawn Opening Zukertort Variation | 71.2 | game |
| 1412 | Mar 29 | Black | Medjdoub86 (1323) | Loss (abandoned) | Pirc Defense Classical Variation | 64.8 | game |
| 1413 | Mar 30 | Black | Grandja (1310) | Loss (abandoned) | Pirc Defense Modern Defense Geller System | 67.1 | game |
The week before was week 62, when I got back over 1300; the next one is week 64. The whole diary is on the journey hub, and if your own Geller or Pirc games keep going wrong, Chess Coach can show you where.
Frequently asked questions
What is the Pirc Defense, Modern Defense Geller System in your games?
It's the name Chess.com gave to three of my Black games that week, all starting 1.e4 g6 2.Nf3, with ...d6 or ...Nf6 and ...Bg7 to follow. In practice it was my usual g6 setup with White's knight on f3 early.
How did the Geller System go for you that week?
Three games, three losses. In game 1405 I was winning by more than five pawns and lost; in game 1413 my queen took three pawns and a rook and was then lost for that rook; game 1410 went wrong with one capture on move 12.
Did you make the same mistake twice?
Yes. In game 1405 the engine's best move was 9...Bxb2, taking a pawn and attacking the undefended rook on a1, the same idea I had missed three weeks earlier in game 1365.