1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.Nf3 d6 5.g3 0-0 6.Bg2 Nbd7 opens the King's Indian Defense: Petrosian Variation, Keres Defense, ECO E67. With 537,147 games on record, the patterns below come from the largest practical sample available.
History and Notable Players
It arises from the King's Indian Defense: Normal Variation, Deferred Fianchetto. On the White side, Zlatko Ilincic (36 games), Istvan Csom (36 games), Josef Jurek (29 games) top the database. Notable Black exponents: Joseph G Gallagher (24 games), Efim Geller (22 games), Miguel Najdorf (22 games).
Performance Across Rating Levels
Popularity and results vary sharply by rating level. At 1200 Elo, the opening shows up in 0.00% of games (10,997 samples). White scores 49.5%, Black 46.7%, draws 3.8%. By 1800, popularity is 0.02% and White's score is 49.6% to Black's 44.9%. At the top end (2500+ Elo), popularity is 0.10% with 9.4% draws — a clear sign of how much theory rules the line at master level. Positions also become less sharp as level rises (sharpness 0.96 → 0.91).
Time Control Patterns
Time control matters here: bullet players reach for this opening more than others. In bullet, it appears in 0.01% of games (358,504); White wins 49.1%. Blitz shows 0.01% adoption across 465,324 games, White scoring 49.5%. In rapid, the share rises to 0.01% — 70,679 games, White 49.4%.
Move Diversity and Theory Depth
What players actually play after the opening moves depends heavily on rating. At 1200 Elo, the top reply is O-O, played 66.7% of the time. There are 3 other moves seeing meaningful share, and 85.8% of games stick to established theory. Entropy: 1.83. By 2500, O-O dominates at 96.8% of replies; only 1 viable alternatives remain and 98.8% of moves are theory. Entropy drops to 0.29. The narrowing is significant — strong players consolidate around a small set of best moves, while amateurs scatter across many plausible-looking options.
Historical Trends
Tracking the King's Indian Defense: Petrosian Variation, Keres Defense year over year shows a clear story. Adoption peaked in 2020 at 0.01% (76,505 games). By 2025 it sits at 0.01% — a 106% shift overall, leaving the line on the rise.
Main Lines and Variations
After 1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.Nf3 d6 5.g3 0-0 6.Bg2 Nbd7, the established follow-ups are:
Each branch leads to a different middlegame character — the resulting pawn structure decides what kind of game you get.
Common Mistakes
- Drifting away from main theory — At 400 Elo, theory adherence sits at 79.2% — versus 95.5% at 2000. The most popular deviation is e4 (played 12.8% of the time at 400, much less so up top). It looks fine but quietly hands the better-prepared side an edge.
- Neglecting development — Extra pawn moves in the opening are tempting, especially when you "know the moves". Developing a piece each turn is the simple correction.
- Letting White own the centre — Hypermodern openings concede central space on purpose, but only if you strike back in time. Delay the counter-blow and you end up squeezed.
Practice on Chessiverse
Ready to try the King's Indian Defense: Petrosian Variation, Keres Defense against a bot? Pick an opponent at your level and play a game.
