losing lines in theory

[Uri Blass]

The folllowing game is from israel human championship between IM and FM
FM Yonathan Uritsky against IM Yeshaayahu Tzidkiya

https://lichess.org/broadcast/israeli-o ... v/5dBQpUk6

I am surprised that strong players choose maybe losing theory lines for black because I thought they are supposed to know theory better and not enter to popular dubious lines.

People tell me that the position is complex and the computer evaluation is meaningless for humans but it does not seem logical for me because strong player may train with engines to study what to do against popular dubious lines.

[pgn][Event "Round 8: Uritsky, Yonathan - Tzidkiya, Yeshaayahu"]
[Site "https://lichess.org/broadcast/israeli-o ... v/5dBQpUk6"]
[Date "2026.01.26"]
[Round "8.2"]
[White "Uritsky, Yonathan"]
[Black "Tzidkiya, Yeshaayahu"]
[Result "1/2-1/2"]
[WhiteElo "2320"]
[WhiteTitle "FM"]
[WhiteFideId "2825651"]
[BlackElo "2470"]
[BlackTitle "IM"]
[BlackFideId "2824493"]
[Variant "Standard"]
[ECO "E69"]
[Opening "King's Indian Defense: Fianchetto Variation, Classical Main Line"]
[UTCDate "2026.01.26"]
[UTCTime "12:00:24"]
[BroadcastName "Israeli Open Chess Championship 2026"]
[BroadcastURL "https://lichess.org/broadcast/israeli-o ... 8/Bz5u8cTv"]
[GameURL "https://lichess.org/broadcast/israeli-o ... v/5dBQpUk6"]

1. Nf3 { [%eval 0.1] [%clk 1:30:44] } 1... Nf6 { [%eval 0.18] [%clk 1:30:54] } 2. g3 { [%eval 0.11] [%clk 1:30:52] } 2... g6 { [%eval 0.21] [%clk 1:31:20] } 3. Bg2 { [%eval 0.13] [%clk 1:31:09] } 3... Bg7 { [%eval 0.18] [%clk 1:31:46] } 4. O-O { [%eval 0.14] [%clk 1:31:28] } 4... O-O { [%eval 0.21] [%clk 1:32:13] } 5. c4 { [%eval 0.1] [%clk 1:31:26] } 5... d6 { [%eval 0.32] [%clk 1:32:24] } 6. d4 { [%eval 0.25] [%clk 1:31:47] } 6... Nbd7 { [%eval 0.54] [%clk 1:32:33] } 7. Nc3 { [%eval 0.43] [%clk 1:31:46] } 7... e5 { [%eval 0.45] [%clk 1:32:58] } 8. h3 { [%eval 0.47] [%clk 1:32:11] } 8... c6 { [%eval 0.69] [%clk 1:32:37] } 9. e4 { [%eval 0.65] [%clk 1:32:34] } 9... exd4 { [%eval 0.76] [%clk 1:31:38] } 10. Nxd4 { [%eval 0.69] [%clk 1:32:57] } 10... Re8 { [%eval 0.68] [%clk 1:32:05] } 11. Be3 { [%eval 0.72] [%clk 1:30:20] } 11... a5 { [%eval 0.74] [%clk 1:29:06] } 12. Qc2 { [%eval 0.72] [%clk 1:29:28] } 12... Nc5 { [%eval 1.03] [%clk 1:28:12] } 13. Rad1 { [%eval 0.86] [%clk 1:27:38] } 13... Nfd7 { [%eval 1.19] [%clk 1:28:37] } 14. b3 { [%eval 1.1] [%clk 1:22:19] } 14... Qb6 { [%eval 1.22] [%clk 1:27:06] } 15. Nde2 { [%eval 0.91] [%clk 1:13:51] } 15... Qb4 { [%eval 0.87] [%clk 1:25:51] } 16. Bd2 { [%eval 1.0] [%clk 1:13:53] } 16... Qb6 { [%eval 0.87] [%clk 1:26:06] } 17. Be3 { [%eval 0.89] [%clk 1:10:13] } 17... Qb4 { [%eval 0.96] [%clk 1:26:32] } 18. Bd2 { [%eval 1.05] [%clk 1:09:07] } 18... Qb6 { [%eval 0.97] [%clk 1:26:57] } 1/2-1/2


[/pgn]


The sides accepted a draw in a line that is maybe winning for white and everything is theory
When I can understand the decision os IM Yeshaayahu Tzidkiya to accept a draw that is good for him to get the GM norm in this tournament(he now needs only a draw in the last round to get a third GM norm and the GM title if I understand correctly) I do not understand the opening choice and how it is possible that strong players play lines that may be losing from preperation.

It seems that everything is theory that players already played in the past.
Stockfish does not like 5...d6 ,6...Nbd7 ,8..c6 12...Nc5 13...Nfd7 and based on stockfish's evaluation white maybe could win after 15.Qb1 that white did not play.

I lopoked also at the statistics
5...d6 ia the most popular move but stockfish does not like it and I wonder if stong players do not use stockfish for their opening choices.
5...c5 or 5...c6 should be more popular for players who use stockfish's analysis to decide about their thoery moves
6...Nbd7 is also very popular and only 6...Nc6 is slightly more popular. but is also not one of the top moves of stockfish.
8...c6 is the most popular move in master games and it is also surprising because stockfish's choice is 8...exd4 or 8...Re8
9...exd4 is at least better than the popular 9....Qb6 and 9...exd4 is probably still a drawing move
11...a5 is less popular relative to 11...Nc5 but still probably not a losing move.

I am not sure if 12...Nc5 or 13...Nfd7 are losing moves.
stockfish after some minutes of search prefer 13...Qe7 that is the most popular move but even 13...Nfd7 is from some human player above 2600 who played it in 2025 and won with black inspite getting a position that may be a losing position.

[pgn][Event "2nd Bundesliga Nord 24-25"]
[Site "Germany GER"]
[Date "2025.01.10"]
[Round "1.3"]
[White "Reuker, Jari"]
[Black "Onischuk, V."]
[Result "0-1"]
[GameId "55vPiuwE"]
[WhiteElo "2401"]
[BlackElo "2609"]
[Variant "Standard"]
[TimeControl "-"]
[ECO "E68"]
[Opening "King's Indian Defense: Fianchetto Variation, Classical Variation"]
[Termination "Unknown"]
[Annotator "lichess.org"]

1. d4 Nf6 2. Nf3 g6 3. g3 Bg7 4. Bg2 O-O 5. c4 d6 6. Nc3 Nbd7 7. O-O e5 8. e4 { E68 King's Indian Defense: Fianchetto Variation, Classical Variation } Re8 9. h3 a5 10. Be3 exd4 11. Nxd4 Nc5 12. Qc2 c6 13. Rad1 Nfd7 14. b3 Qb6 15. Rd2 a4 16. Rb1 axb3 17. axb3 Qc7 18. b4 Ne6 19. Ndb5 cxb5 20. Nxb5 Qb8 21. Nxd6 Rd8 22. f4 Nf6 23. c5 Ne8 24. f5 Nxd6 25. fxe6 Bxe6 26. cxd6 Rxd6 27. Bf4 Qa7+ 28. Kh2 Rc6 29. Qd1 h5 30. e5 Rc4 31. b5 Qb8 32. Qe2 h4 33. Rd6 Qc7 34. b6 Qe7 35. Rbd1 Rac8 36. Qe3 Rc3 37. R1d3 Rxd3 38. Rxd3 Bf5 39. Rd6 g5 40. Bxg5 Qxe5 41. Rd5 hxg3+ 42. Qxg3 Qe6 43. Bf4 Bg6 44. Rd6 Qe7 45. Qg4 Rc2 46. Rd7 Qf6 47. Rd6 Qe7 48. Qd1 Qe4 49. Rd8+ Kh7 50. Bd2 Be5+ 51. Kh1 Qh4 52. Rd5 Qg3 53. Kg1 Bf4 54. Qe2 Qh2+ 55. Kf1 Bxd2 56. Rxd2 Qf4+ 57. Ke1 Rc1+ 58. Rd1 Qg3+ 59. Kf1 Qf4+ 60. Ke1 Qh4+ 61. Kf1 Qf6+ 62. Ke1 Qc3+ 63. Kf2 Rc2 { Black wins. } 0-1[/pgn]

[FireDragon761138]

I'm dubious of engines' abilities to evaluate openings. Opening theory is all about creating practical chances. Stockfish dislikes alot of openings that actually do well at various levels of play, just by judging off game databases.

[Eelco de Groot]

Also, I think Vladimir Onischuk is not just any GM and if he wins this, with Black, I am betting he was not out of theory yet with 13...Nfd7. But (Uri) you should ask some people with big openingbooks etc or chessplayers here. That's not me. I asked Rebel EAS 3.0 and he thinks White is about 1.4 up after this move, I played a few moves along Rebel's line and then further with Sun SE, the position after Rebel's 17. f4 At least two defences seem playable here says SUN, 17...h5 and 17...Nf8.

If you think you can still search for chessgames with Google: apparently not me, it only found a little information about Jari, the White player. No games between these two players. A bit shocking. Ten years ago, you could find pretty much any game, ever played with just one Google click.

[pgn]
[Event "2nd Bundesliga Nord 24-25"]
[Site "Germany GER"]
[Date "2025.01.10"]
[Round "1.3"]
[White "Reuker, Jari"]
[Black "Onischuk, V."]
[WhiteElo "2401"]
[BlackElo "2609"]
[ECO "E68"]
[Result "0-1"]

1. d4 Nf6 2. Nf3 g6 3. g3 Bg7 4. Bg2 O-O 5. c4 d6 6. Nc3
Nbd7 7. O-O e5 8. e4 {E68 King's Indian Defense: Fianchetto
Variation, Classical Variation } Re8 9. h3 a5 10. Be3 exd4
11. Nxd4 Nc5 12. Qc2 c6 13. Rad1 Nfd7 14. b3 (14. b3
{+1.46/33 24:46m} Qb6 15. Qb1 Qd8 16. Rfe1 Qe7 17. f4 Nf8
18. Kh2 Bd7 19. Qc1 Rad8 20. Nf3 Bc8 21. Bf2 Nfe6 22. Qc2
Qc7 23. Qd2 Nd7 24. Re2 Nec5 25. Nd4 Nf6 26. Kg1 h6 27. a3
{This was Rebel's line. No backchecking yet.}) 14... Qb6
15. Rd2 a4 16. Rb1 axb3 17. axb3 Qc7 18. b4 Ne6 19. Ndb5
cxb5 20. Nxb5 Qb8 21. Nxd6 Rd8 22. f4 Nf6 23. c5 Ne8 24. f5
Nxd6 25. fxe6 Bxe6 26. cxd6 Rxd6 27. Bf4 Qa7+ 28. Kh2 Rc6
29. Qd1 h5 30. e5 Rc4 31. b5 Qb8 32. Qe2 h4 33. Rd6 Qc7
34. b6 Qe7 35. Rbd1 Rac8 36. Qe3 Rc3 37. R1d3 Rxd3 38. Rxd3
Bf5 39. Rd6 g5 40. Bxg5 Qxe5 41. Rd5 hxg3+ 42. Qxg3 Qe6
43. Bf4 Bg6 44. Rd6 Qe7 45. Qg4 Rc2 46. Rd7 Qf6 47. Rd6 Qe7
48. Qd1 Qe4 49. Rd8+ Kh7 50. Bd2 Be5+ 51. Kh1 Qh4 52. Rd5
Qg3 53. Kg1 Bf4 54. Qe2 Qh2+ 55. Kf1 Bxd2 56. Rxd2 Qf4+
57. Ke1 Rc1+ 58. Rd1 Qg3+ 59. Kf1 Qf4+ 60. Ke1 Qh4+ 61. Kf1
Qf6+ 62. Ke1 Qc3+ 63. Kf2 Rc2 {Black wins. } 0-1
[/pgn]

After 17.f4:

[d]r1b1r1k1/1p1nqpbp/2pp2p1/p1n5/2PNPP2/1PN1B1PP/P5B1/1Q1RR1K1 b - -

Engine: Sun SE (1024 MB)
gemaakt door E. Nemeth, the Stockfish developers (se

38 63:28 +1.10 17...h5 18.Dc2 Pf8 19.Kh2 Pfe6
20.Df2 Pxd4 21.Lxd4 Lxd4 22.Txd4 f5
23.Ted1 Pxe4 24.Pxe4 fxe4 25.Lxe4 c5
26.Ld5+ Kf8 27.Te4 Df6 28.Txe8+ Kxe8
29.Le4 Kf7 30.Dd2 a4 (1.855.527.550) 487

38 63:28 +1.14 17...Pf8 18.Dc2 h5 19.Kh2 Pfe6
20.Pde2 Pf8 21.Lf2 Ld7 22.Pg1 Tad8
23.Pf3 Lc8 24.Ph4 Lf6 25.a3 Lxh4
26.gxh4 f5 27.Db2 Dc7 28.b4 Pxe4
29.Lxe4 fxe4 30.Pxe4 Ph7 (1.855.527.550) 487

38 63:28 +1.35 17...Df8 18.Lf2 h5 19.Dc2 Pb8 20.Pf3 Pba6
21.Ld4 Pb4 22.Db1 Lxd4+ 23.Pxd4 Pba6
24.Pf3 Le6 25.Ph4 Dg7 26.Pe2 Tad8
27.Kh2 Lc8 28.a3 Df6 29.Dc2 Pd7
30.Pd4 Pac5 (1.855.527.550) 487

38 63:28 +1.36 17...Pa6 18.Lf2 Pdc5 19.a3 h5 20.Kh2 Pc7
21.Dc2 Df8 22.b4 axb4 23.axb4 Pd7
24.b5 c5 25.Pb3 h4 26.gxh4 Ta3 27.e5 Pe6
28.Db2 Txb3 29.Dxb3 dxe5 30.Pe4 Pd4 (1.855.527.550) 487

38 63:28 +1.41 17...Pb8 18.Lf2 Pba6 19.a3 Pc7
20.Dc2 P7a6 21.Td2 Ld7 22.Db2 h5
23.b4 Pe6 24.Pf3 axb4 25.axb4 Pf8
26.Ted1 Le6 27.Txd6 Lxc4 28.e5 Pc7
29.h4 f6 30.exf6 Lxf6 (1.855.527.550) 487
______________________________________________________

39 110:16 +1.08 17...Pf8 18.Dc2 h5 19.Kh2 Pfe6
20.Df2 Pxd4 21.Lxd4 Lxd4 22.Txd4 f5
23.Ted1 Pxe4 24.Pxe4 fxe4 25.Lxe4 c5
26.Ld5+ Kf8 27.Te4 Df6 28.Txe8+ Kxe8
29.Le4 Kf7 30.Dd2 a4 (2.905.625.650) 439

39 110:16 +1.10 17...h5 18.Dc2 Pf8 19.Kh2 Pfe6
20.Df2 Pxd4 21.Lxd4 Lxd4 22.Txd4 f5
23.Ted1 Pxe4 24.Pxe4 fxe4 25.Lxe4 c5
26.Ld5+ Kf8 27.Te4 Df6 28.Txe8+ Kxe8
29.Le4 Kf7 30.Dd2 a4 (2.905.625.650) 439

39 110:16 +1.34 17...Df8 18.Lf2 h5 19.Kh2 Pf6 20.Dc2 Ph7
21.Lf3 Pf6 22.Dd2 Td8 23.Kg2 Ld7
24.Pc2 Le6 25.Kh2 Ph7 26.Pd4 Te8
27.Kg2 Ld7 28.g4 g5 29.fxg5 hxg4
30.hxg4 Pe6 (2.905.625.650) 439

39 110:16 +1.36 17...Pa6 18.Lf2 h5 19.Kh2 Pac5
20.Dc1 Df8 21.Dc2 Pb8 22.Tf1 De7
23.a3 Pbd7 24.Tfe1 Df6 25.b4 axb4
26.axb4 Pe6 27.Pxe6 Txe6 28.Ld4 De7
29.Lxg7 Kxg7 30.Pe2 Df8 (2.905.625.650) 439

39 110:16 +1.40 17...Pb8 18.Lf2 Pba6 19.a3 Pc7
20.Dc2 P7a6 21.g4 Ld7 22.Pce2 Df8
23.f5 b5 24.Pg3 b4 25.a4 Tad8 26.h4 Le5
27.Pf3 De7 28.g5 Lc8 29.Pe2 Pd7
30.Ped4 Lb7 (2.905.625.650) 439
______________________________________________________

40 133:02 +1.11 17...Pf8 18.Dc2 h5 19.Lf2 Ld7 20.Kh1 Pfe6
21.Pf3 Tad8 22.a3 Df6 23.Te3 Pf8
24.b4 axb4 25.axb4 Pa6 26.Db1 Le6
27.Lf1 De7 28.Pa4 b5 29.cxb5 cxb5
30.Lxb5 Ld7 (3.364.852.980) 421

40 133:02 +1.17 17...h5 18.Dc2 Pf8 19.Lf2 Ld7 20.Kh1 Pfe6
21.Pf3 Tad8 22.a3 Df6 23.Te3 Pf8
24.Kh2 Le6 25.Db1 Lc8 26.b4 axb4
27.axb4 Pce6 28.Ted3 g5 29.f5 g4
30.Pg1 Pd7 (3.364.852.980) 421

[Geezer one]

The game looks completely normal to me. Engines famously don't like the King's Indian defense, but it is fine for experienced people.

[Ajedrecista]

Hello Eelco:

[...]

If you think you can still search for chessgames with Google: apparently not me, it only found a little information about Jari, the White player. No games between these two players. A bit shocking. Ten years ago, you could find pretty much any game, ever played with just one Google click.

[...]

I found a likely complete profile of Reuker at ChessTempo:

Jari Reuker - Games

Where the only recorded game against Onischuk is this one. I am also very pleased about the new ChessTempo opening explorer, now that the former 10-move limit has been removed.

Regarding Google searches, I find a similar behaviour for some time now, in my case when I search football results of live matches, for example Real Madrid: formerly, Google displayed updatable info about the current elapsed time, the current score and who and when scored the goals; nowadays, I only get the score (not updatable anymore if someone scores), if the match is in the first half, half-time, second half or finished (not updatable again), and no info about scorers and when.

Regards from Spain.

Ajedrecista.

[cpeters]

Slightly ot, but I (and every ex-pioneer from the former eastern block whom I talked to in this matter) do find the transcription (Onischuk) very strange - and it's FIDE/official.

greetings

[Stephen Ham]

Hi Christian,

So how should Onischuk be transcribed?

-Steve-