TalkChess.com

Lyudmil Tsvetkov wrote:[d]r1br2k1/1p1n1ppp/p1p1pn2/2Pp4/3P4/qPN1PN2/2BB1PPP/3Q1RK1 b - - 0 3

anyone having a definitive conclusion on this position, arising after a3 Ba3 Ra3 Qa3 c5:

[d]r1br2k1/1p1n1ppp/p1p1pn2/2Pp4/3P4/qPN1PN2/2BB1PPP/3Q1RK1 b - - 0 3

is there a win?

for me, it is an obvious draw.

b6 might be losing, but both a5 and e5 should give black a straightforward draw, too much tactics for any human to see, but SF claims both lines are 0.0.

i have no idea how you get that output when a5 loses on the spot. maybe you mean b5 which computers initially seem to like...they end up switching to b6. SF eval is not even close to 0.0 for either b6 or e5 except at low depths.

yanquis1972 wrote:

analyses on youtube and similar channels are usually low quality.

youre really something else, man.

svidler is a super-GM who makes dutiful use of engines to check his analysis. his analysis of the game is top-notch given the circumstances (he'd just covered the entire round & had to post it that same day), but he admits its incomplete.

i don't recall him saying Bxa3 loses but i could be wrong. he did say something along the lines of Bd6 being workable & objectively best but restrained & passive.

a3 is a brilliant move any way you can possibly look at it. computers crunch the permutations & evaluate it as the best option in the position. it was essentially a novelty that took carlsen completely offguard. c5 might give winning chances but its not the way to beat carlsen...& its much less likely to win against any other human you have the chance to spring it on.

no one has said a3 wins on the spot, but that seems to be your criteria w/r/t its merits. that, or you truly believe yourself to have more insight into the position in 5 minutes than aronion, carlsen, svidler, top-end machines, & everyone else in the world. either way you walk alone here. i suppose you find that noble, or something.

you did not give a single proof the posted position is not draw, not a single one, just random, partly hot-blooded remarks, so I suppose I am generally right.

at the end of the day, does it really matter who tells the truth: Aronian, Svidler or myself? what is important is that it is the truth.

is there any restriction that a person who has not been playing for 12 or so years tournament chess could not have added some 500-600 elo in the process, objectively, without that being registered in any official way?

as far as I know, FIDE does not forbid that, as well as none other organisation or rule I know of.

pity again no one analyses to see the objective state of affairs, and instead incriminations start pouring down again...

yanquis1972 wrote:

Lyudmil Tsvetkov wrote:[d]r1br2k1/1p1n1ppp/p1p1pn2/2Pp4/3P4/qPN1PN2/2BB1PPP/3Q1RK1 b - - 0 3

anyone having a definitive conclusion on this position, arising after a3 Ba3 Ra3 Qa3 c5:

[d]r1br2k1/1p1n1ppp/p1p1pn2/2Pp4/3P4/qPN1PN2/2BB1PPP/3Q1RK1 b - - 0 3

is there a win?

for me, it is an obvious draw.

b6 might be losing, but both a5 and e5 should give black a straightforward draw, too much tactics for any human to see, but SF claims both lines are 0.0.

i have no idea how you get that output when a5 loses on the spot. maybe you mean b5 which computers initially seem to like...they end up switching to b6. SF eval is not even close to 0.0 for either b6 or e5 except at low depths.

I told you I did not analyse the position, but just quickly let SF run on it, and it said 0.0 for both a5 and e5. maybe one of them is losing at bigger depths, I was using less than a second search.

in any case, if SF claims there are plentiful 0.0 draws in a purely tactical position, and I myself would not think white has a forced win, I am inclined to believe it.

again, any definitive proof the position is won?
what are your/engine highest scores after analysis?
higher, or lower than 50cps?

because, in the c5 line, white should get close to that one advantage, so, if score is lower than that, then c5 will certainly be the better move.

ok, I analysed it, and there is this long, and completely forced line, ending with a draw:

[pgn][Event "Blitz 1m"]
[Site "Microsoft"]
[Date "2017.06.12"]
[Round "?"]
[White "Stockfish 8 64 POPCNT"]
[Black "myself, owner"]
[Result "*"]
[Annotator "owner"]
[SetUp "1"]
[FEN "r1br2k1/1p1n1ppp/p1p1pn2/2Pp4/3P4/qPN1PN2/2BB1PPP/3Q1RK1 b - - 0 3"]
[PlyCount "29"]
[TimeControl "60"]

{512MB, OWNER-PC} 3... Ne4 {0} 4. Bxe4 {1.20/22 6} dxe4 {2} 5. Ng5 {0.94/22 0}
h6 {2} 6. Ngxe4 {1.39/19 1} a5 {4} 7. Qc2 {1.16/24 2} a4 {1.08/25 1} 8. bxa4 {
0.97/23 1} Qb4 {1.09/25 1} 9. Nd6 {1.21/24 0} Qa5 {1.10/25 0} 10. Nd5 {1.14/25
1} Qxa4 {1.19/26 0} 11. Ne7+ {1.02/26 0} Kf8 {1.02/26 0} 12. Qxa4 {1.00/23 0}
Rxa4 {1.10/27 0} 13. Nexc8 {1.04/25 0} Ra8 {0.98/22 0} 14. Rb1 {1.10/25 0}
Rdxc8 {1.02/24 0} 15. Nxc8 {1.03/23 0} Rxc8 {0.88/22 0} 16. Rxb7 {0.94/21 0}
Ra8 {1.08/23 0} 17. Kf1 {0.95/23 0} Ke7 {0.98/20 0} *

[/pgn]

completely forced, completely drawn, white can not win that, too few material.

capturing with the knight on e4 first is even worse, complete equality.

so, seemingly, a3 indeed only leads to a quick and straightforward draw.

why choose a straightforward draw in 15 moves in a forced line, when you have much better?

therefore, I have evry right in the world to suppose c5 is much better than a3.

Intel E5-2666 v3, 8 threads, 16 GB Hash respectively Intel E5-2686 v4, 32 Threads, 32 GB Hash

Some Stockfish analyzes after 12... Qxa3 13.c5:

[d]r1br2k1/1p1n1ppp/p1p1pn2/2Pp4/3P4/qPN1PN2/2BB1PPP/3Q1RK1 b - - 0 13

Stockfish 110617 64 BMI2:
47 3:29:58 179.445.570k 14.244k -0,72 13. ... b5 14.Nb1 Qa2 15.Nc3 Qa3 16.Ne1 a5 17.Qb1 a4 18.bxa4 Rb8 19.Nxb5 cxb5 20.Bb4 bxa4 21.Bxa3 Rxb1 22.Bxb1 Ne4 23.Bb4 Ba6 24.Bd3 Nb8 25.Ba5 Rc8 26.f3 Sf6 27.Bc3 e5 28.Rf2 exd4 29.Bxd4 Nfd7 30.Rb2 a3 31.Ra2 Nxc5 32.Bxa6 Ncxa6 33.Rxa3 Nb4 34.Ra8 f6 35.Bc5 N4c6 36.Nd3 Kf7 37.Nf4 Ne5 38.Bd6 Rc1+ 39.Kf2 Nbd7 40.Bxe5 fxe5 41.Nxd5 Rc2+ 42.Kg3 Nf6


[d]r1br2k1/1p1n1ppp/2p1pn2/p1Pp4/3P4/qPN1PN2/2BB1PPP/3Q1RK1 w - - 0 14

Stockfish 120617 64 BMI2:
43/62 04:15 13.081.260k 51.127k +2,48 14.Bc1 Qa1 15.Qd2 b6 16.cxb6 Ba6 17.Re1 h6 18.Bb1 Ne4 19.Bxe4 dxe4 20.Bb2 Qxb2 21.Qxb2 exf3 22.gxf3 Nxb6 23.Ra1 Bb5 24.Ne4 Rd7 25.Qc3 Rda7 26.Kg2 Nd7 27.Qc2 Kh8 28.Kh3 Kg8 29.Qc3 Kf8 30.Kg3 Kg8 31.Kg2 Ra6 32.Qd2 a4 33.bxa4 Rxa4 34.Rxa4 Rxa4 35.Qc3 Nb6 36.Qc5 Nd5 37.Qd6 Ra2 38.Qb8+ Kh7 39.Kg3 Ra1 40.h3 Ne7 41.Nd6 Nf5+ 42.Nxf5 exf5 43.Qe8 Kg6


[d]r1br2k1/3n1ppp/ppp1pn2/2Pp4/3P4/qPN1PN2/2BB1PPP/3Q1RK1 w - - 0 14

Stockfish 120617 64 BMI2:
48/70 25:32 75.529.857k 49.292k +0,59 14.Nb1 Qa2 15.Qc1 bxc5 16.Nc3 Qa5 17.Nxd5 Qb5 18.Nc7 Qb7 19.Nxa8 Qxa8 20.Ba5 Re8 21.dxc5 Nxc5 22.Bxh7+ Nxh7 23.Qxc5 Nf6 24.b4 Qb8 25.Ne5 Nd7 26.Nxd7 Bxd7 27.Rd1 Qb7 28.h3 e5 29.Rd6 Bf5 30.Kh2 Qb5 31.Rxc6 Qe2 32.f3 Bd7 33.Rd6 Bb5 34.e4 Qc4 35.Bb6 f6 36.Qd5+ Qxd5 37.exd5 Re7 38.Rd8+ Re8 39.Rxe8+ Bxe8 40.Kg3 Kf7 41.f4 exf4+ 42.Kxf4 g5+ 43.Kf5 Bd7+ 44.Ke4 Bb5 45.Be3 Bc4 46.Bf2 Bf1 47.Kf3 Bb5


[d]r1br2k1/1p1n1ppp/p1p2n2/2Ppp3/3P4/qPN1PN2/2BB1PPP/3Q1RK1 w - - 0 14

Stockfish 120617 64 BMI2:
48/79 20:37 61.006.778k 49.288k +1,29 14.Nb1 Qa2 15.Bb4 Ne4 16.Bxe4 dxe4 17.Nc3 Qb2 18.Na4 Qa2 19.Nd2 exd4 20.exd4 Ne5 21.dxe5 Be6 22.Nb6 Rab8 23.Qe1 Rd3 24.Nxe4 Qxb3 25.Nd6 Qb2 26.Ba5 Qd4 27.Qc1 h6 28.Re1 Rb3 29.Bd2 Qb2 30.h3 h5 31.Rd1 Qxc1 32.Rxc1 g6 33.f4 Kg7 34.Bc3 Ra3 35.Bd4 Rb3 36.Kf2 Ra3 37.Rd1 Ra2+ 38.Kg1 Kh7 39.Re1 Rd2 40.Be3 Rb2 41.Kf1 Ra2 42.Re2 Ra3 43.Kf2 a5 44.Kf3


[d]r1br2k1/3n1ppp/ppp1pn2/2Pp4/1P1P4/q1N1PN2/2BB1PPP/3Q1RK1 b - - 0 14

Stockfish 120617 64 BMI2:
46/67 13:04 38.774.110k 49.401k 0,00 14. ... Qb2 15.Na4 Qa3 16.cxb6 Ne4 17.Nc5 Nxd2 18.Qxd2 Nxb6 19.Rb1 Nd7 20.Rb3 Qa1+ 21.Rb1 Qa3


[d]r1br2k1/3n1ppp/ppp1p3/2Pp4/1P1Pn3/q1N1PN2/2BB1PPP/3Q1RK1 w - - 1 15

Stockfish 120617 64 BMI2:
47/57 06:25 19.620.051k 50.854k 0,00 15.Nxe4 dxe4 16.Bxe4 bxc5 17.bxc5 a5 18.Bxc6 Ra7 19.Qa1 Qxa1 20.Rxa1 Nb8 21.Bb5 a4 22.Rxa4 Rxa4 23.Bxa4 Bd7 24.Bb3 Bb5 25.Ba5 Rc8 26.Nd2 Nc6 27.Bc3 Ra8 28.f3 Na5 29.Ba2 Nc6 30.Bb3


[d]r1br2k1/3n1ppp/p1p1p3/2p5/1P1PB3/q3PN2/3B1PPP/3Q1RK1 w - - 0 17

Stockfish 120617 64 BMI2:
47/65 16:37 52.234.220k 52.347k -0,08 17.bxc5 a5 18.Bxc6 Ra7 19.Qc2 Ba6 20.Rc1 Nb8 21.Ba4 Bd3 22.Qd1 Be2 23.Qc2 Na6 24.Ne5 Nb4 25.Bxb4 axb4 26.Bb3 Rda8 27.Re1 Bb5 28.g3 Ba4 29.Rb1 Bxb3 30.Rxb3 Qa1+ 31.Kg2 Ra2 32.Qb1 Qxb1 33.Rxb1 Rb8 34.Nc6 Rb5 35.Rxb4 Rxb4 36.Nxb4 Ra4 37.Nd3 Ra3 38.Nf4 Kf8 39.h4 e5 40.dxe5 Rc3 41.c6 Rxc6 42.Ne2 Rc5 43.f4 h5 44.Kf3 g6 45.Nd4 Ke7


[d]1rbr2k1/3n1ppp/ppp1p3/2P5/1P1PB3/q3PN2/3B1PPP/3Q1RK1 w - - 1 17

Stockfish 120617 64 BMI2:
47/76 34:32 107.409.842k 51.829k +1,36 17.Qc2 f5 18.Bc3 fxe4 19.Nd2 bxc5 20.Nb1 Qxb4 21.Bxb4 cxb4 22.Nd2 Nf6 23.Nb3 Rf8 24.h3 h6 25.Qxc6 Nd5 26.Ra1 Rb6 27.Qc4 Bb7 28.Nc5 Re8 29.Ra5 Bc6 30.Qc1 Bb5 31.Qb2 Bd3 32.Kh2 Nc3 33.f3 Re7 34.Qb3 Kh8 35.h4 Re8 36.fxe4 Bxe4 37.Rxa6 Bd5 38.Qc2 Rbb8 39.Ra1 Rec8 40.Qg6 Rg8 41.Qd3 Ra8 42.Rxa8 Rxa8 43.Kg3 Rb8


[d]1rbr2k1/3n2pp/ppp1p3/2P5/1P1Pp3/q1B1P3/2QN1PPP/5RK1 b - - 1 19

Stockfish 120617 64 BMI2:
45/74 12:41 40.466.537k 53.118k -1,38 19. ... bxc5 20.Nb1 Qxb4 21.Bxb4 cxb4 22.Nd2 Nf6 23.h4 Bd7 24.Ra1 Rb6 25.Qc5 Rdb8 26.Nc4 R6b7 27.Rxa6 b3 28.Ra1 b2 29.Rb1 Rb3 30.Qa7 R3b4 31.Qa2 c5 32.Rxb2 Rxb2 33.Nxb2 cxd4 34.exd4 Rb4 35.Nc4 h6 36.Ne5 Rxd4 37.Nxd7 Rd1+ 38.Kh2 Rxd7 39.Qxe6+ Kh8 40.g3 Rd5 41.Kg2 Kh7 42.Kf1 Kg6 43.Ke2 Kh7 44.Qf7 Rb5 45.Kf1 Rb1+ 46.Kg2 Rd1 47.Qc4 Rd5 48.Qb4 Rd3 49.Qb2 Rd8 50.Qc2 Rd5 51.Qc3 Kg6

to correct one opinion of mine: capturing on a3 is the only alternative for black, as otherwise, after Bd6 c5 (too many tempos lost) Bc7 e4:

[d]r1br2k1/1pbnqppp/p1p1pn2/2Pp4/3PP3/PPN2N2/2BB1PPP/R2Q1RK1 b - - 0 3

this is certainly won for white.

no time to do any deeper check, but evaluation is simply winning.

so, seemingly, Carlsen chose the right move by capturing, and what engines suggest to retreat on d6, is completely lost.

again, for me, main line remains c5 e5 a3 Ba3 Ba3 e4 Nd2, this is the deepst and most intriguing line.

Lyudmil Tsvetkov wrote:ok, I analysed it, and there is this long, and completely forced line, ending with a draw:

[pgn][Event "Blitz 1m"]
[Site "Microsoft"]
[Date "2017.06.12"]
[Round "?"]
[White "Stockfish 8 64 POPCNT"]
[Black "myself, owner"]
[Result "*"]
[Annotator "owner"]
[SetUp "1"]
[FEN "r1br2k1/1p1n1ppp/p1p1pn2/2Pp4/3P4/qPN1PN2/2BB1PPP/3Q1RK1 b - - 0 3"]
[PlyCount "29"]
[TimeControl "60"]

{512MB, OWNER-PC} 3... Ne4 {0} 4. Bxe4 {1.20/22 6} dxe4 {2} 5. Ng5 {0.94/22 0}
h6 {2} 6. Ngxe4 {1.39/19 1} a5 {4} 7. Qc2 {1.16/24 2} a4 {1.08/25 1} 8. bxa4 {
0.97/23 1} Qb4 {1.09/25 1} 9. Nd6 {1.21/24 0} Qa5 {1.10/25 0} 10. Nd5 {1.14/25
1} Qxa4 {1.19/26 0} 11. Ne7+ {1.02/26 0} Kf8 {1.02/26 0} 12. Qxa4 {1.00/23 0}
Rxa4 {1.10/27 0} 13. Nexc8 {1.04/25 0} Ra8 {0.98/22 0} 14. Rb1 {1.10/25 0}
Rdxc8 {1.02/24 0} 15. Nxc8 {1.03/23 0} Rxc8 {0.88/22 0} 16. Rxb7 {0.94/21 0}
Ra8 {1.08/23 0} 17. Kf1 {0.95/23 0} Ke7 {0.98/20 0} *

[/pgn]

completely forced, completely drawn, white can not win that, too few material.

capturing with the knight on e4 first is even worse, complete equality.

so, seemingly, a3 indeed only leads to a quick and straightforward draw.

why choose a straightforward draw in 15 moves in a forced line, when you have much better?

therefore, I have evry right in the world to suppose c5 is much better than a3.

looks completely lost to me and near zugzwang after 18. e4! Kd8 19. Bf4

There is a video about this game on YouTube with this position at 5:55. Peter Svidler gives some analysis

https://www.youtube.com/watch?v=5XBcbpc7ays

yanquis1972 wrote:well, thats rather astounding...

where did you find a june 11 build of SF? official site still lists june 6 as latest update available.

Not really , if you build it yourself it always takes the current date as the build date. That's true even if you take some random source from the past. The last commit was 6 days ago - so building it yourself today is the same as June 6th anyways. One advantage of building SF yourself and that it's usually a tad faster since the compiler will optimize SF for your machine if using pgo and -mutune= native flag as an example.

Final position of your analyzes:
[d]r7/1R1nkpp1/2p1p2p/2P5/3P4/4P3/3B1PPP/5K2 w - - 3 28
It's a relatively forced line for white but Stockfish shows nearly +2 in the final position of your analyzes.

49/89 1:02:29 221.124.736k 58.975k +1,84 28.e4 Ra1+ 29.Ke2 Ra4 30.Kd3 f6 31.Rc7 Ra6 32.Bf4 e5 33.dxe5 Kd8 34.exf6 gxf6 35.Ke2 Nxc5 36.Rh7 Nxe4 37.Rxh6 Nc3+ 38.Kf3 Ke7 39.Be3 Ra4 40.h4 Nd5 41.Rh7+ Ke6 42.g3 Ra3 43.Ra7 Rc3 44.h5 Nxe3 45.fxe3 Rc1 46.g4 Rh1 47.Kg2 Rh4 48.Kg3 Rh1 49.Kf4 Rf1+ 50.Ke4 Rh1 51.Rc7 Kd6 52.Rc8 Kd7 53.Rf8 Ke6 54.Rg8 c5 55.Re8+ Kf7 56.Rc8 Rg1 57.Kf4 Rf1+ 58.Kg3 Rc1 59.Rc7+ Ke6 60.h6 Rh1 61.h7 Kd5