Mar1112   notyetagm: Damn, Carlsen was 10(!!) when he played this game. 

Jun2114
  wwall: 58...Rxa6? is the losing move. Black should be able to draw after 58...f2 59.Rg8+ Kh5 60.Rf8 Rxa6 61.Kc5 Ra5+ 62.Kc6 Ra2 63.b7 Rc2+ and 64...Rb2 

Jul3014   Ediciones3jaques: Esta partida habrá perdido por tiempo, porque no veo como el negro se impone al final, eso sí hay tablas luego de 58...f2!! 

Apr1315   Cactusjuice: I think 11. Qa5 was a bad move 

Jun1316   Sergash: This was the strongest player Carlsen had ever faced in a tournament game at that time. More than 500 rating points were separating the 2 players! Giorgi Kacheishvili (2583), born in Georgia (formerly part of USSR) in 1977 (he was 23 or 24 years old at the time), international master in 1994, grandmaster since 1997, Georgia Champion in 2006 and still rated 2589 early 2016 (his peak being 2613 in 2009); Magnus Carlsen (2072 having lost 12 points from his personnal record of 2084 since last tournament), no title, born in 1990 (he was 10 years old). A rating difference that high gives a 5% chance of winning according to Arpad Elo's formula (see http://www.3dkingdoms.com/chess/elo...). I went through this game with the program Komodo 9.42 64 bits. Not only Carlsen did get the better position from the opening, as Black, but maintained equality at least until move 50! <2...e6> Earlier that same year, Carlsen had played 2...c5 here and won the game  see P Skovgaard vs Carlsen, 2001 <10...Nc6N> This was apparently the theoritical novelty of this game, which doen't seem to improve anything. Before that, there had been 10...h6 11.Bxf6 (or 11.Bh4 a5 12.f3 Re8 (by transposition) Bernd Kohlweyer (2435) vs. Sergey Kishnev (2470), BL2West (league) 9495 (Germany) 1995, round 7  St. Ingbert against Gelsenkirchen, 01) Qxf6 12.Qxf6 gxf6 JeanLuc Chabanon (2445) vs. Etienne Bacrot (2300), Cannes Hotel Mattinez (France) 1995, round 10, 10. Also, there was 10...d6 11.Bxf6 Qxf6 12.Qxf6 gxf6 13.000 Rd8 14.Nc3 a6 = / Irina Krush (2469) vs. Robert Hess (2601), Chicago Open (USA) 2011 in Wheeling, 01. 

Jun1316   Sergash: <11...Qa5!> To answer Cactusjuice ( <Apr1315 Cactusjuice: I think 11. Qa5 was a bad move> ), according to the computer this is a strong move. Forcing the exchange of the queens helps Black obtain equality. click for larger viewDIAGRAM <15.000?! Kf8! / = > This is the first error in the game, and made by the stronger player, possibly by analogy with other variations where this move is common. But here it is inferior and Kacheishvili probably did not pay any particular attention to Carlsen's strong reply. White should have captured the pawn: 15.Nxf6+ Kg7 16.Nh5+! and now: A) 16...Kf8 17.b4! Nc6! 18.b5 Na5 (or 18...Ne5 19.f4! Ng4! 20.Ke2! followed with h2h3) 19.f3 or 19.Rc1 B) 16...Kh6! 17.Nf4! Rdc8!
Or 15.b4 Nc6! 16.b5! and now:
A) 16...Ne5 17.Nxf6+ (or 17.f4 Ng4! 18.Ke2 ) Kg7 18.Nh5+ Kh6 (or 18...Kf8 transposing) 19.Nf4 B) 16...Na5 17.Nxf6+ Kg7 18.Nh5+ Kh6 19.Nf4 or 19.Ng3 

Jun1316   Sergash: <17.Nh5?!> It is surprising that a player as strong as Kacheishvili playing White is being dominated in the opening by an untitled child! He should simply had played 17.Nxh7! and now what would Black play? A) 17...Rg8 (idea: Rg8g7) 18.b4! Nc6 (or 18...cxb4 19.axb4 Nc6 20.h4 transposing) 19.h4! cxb4 (or 19...Ne5 20.Ng5 cxb4 21.axb4 Nxc4! 22.Bxc4 Rcg8! ) 20.axb4 Nxb4 21.Ng5 B) 17...f6 (idea Rd8h8) 18.Bd3 Rh8! (if 18...Rg8 19.Rhg1! =) 19.h4 C) 17...Rab8 18.Bd3
<17...Ba6> Stronger is 17...Rab8! 18.Nf4! Ba6 (or 18...Be4 19.Bd3! Bxd3 20.Nxd3 Nxc4 ) 19.Bd3 Bxc4 20.Bxh7 Rb6 or 20...Rb7 in both cases planning to follow with Rdb8. <18.Nf4 Bxc4 19.Bxc4 Nxc4 / > Better seems 18.Kb1! Bxc4 (or 18...Rab8 19.Rc1 Rb3! ) 19.Bxc4 Nxc4 20.Rc1 d5 <22...Na5> The idea is, of course, that if 23.Rxc5?? Nb3+ and Nxc5. <24...Nb7?! 25.e4! = > Only ONE weaker move and all this nice advantage Carlsen had managed to built goes up in smoke... 24...Rd6! 25.Ka2 Rb3 / click for larger viewDIAGRAM. <27.Nc3> Though complex to calculate, especially near move 30 when one is getting short on time on the clock, somewhat better is 27.Nd4! threatening Nd4c6+. Black has 2 main ways of replying to this: A) 27...Ke8 28.Rxc4! e5! 29.Re1! Rxb2+! 30.Kxb2 Nd3+ 31.Kb3 (or 31.Kc3 Nxe1 32.Nc6 (or 32.Nf5 Nxg2 33.Rc7! =) Nxg2 33.a4 =) Nxe1 32.Nc6 (or 32.Nf5 Nxg2 33.Rc7! =) Nxg2 33.Nxa7 = B) 27...Kf6 28.Rxc4! e5! 29.Rdc1! Nd3! 30.Rc6+! Ke7 (or 30...Kg5 31.Nf3+ Kf5 32.Rc1c2! = / ; but not 30...Kg7? 31.Nf5+! Kf8 32.R1c2! ) 31.Nf5+! Kd7! 32.R1c2! = 

Jun1316   Sergash: <27...Rg5 28.f4 = > Now the position is completely equal and steering towards a draw. Somewhat better was 27...Rd3! 28.Re1! Rdb8 = / <30...Nd3?!> Weak. Getting short on time? 30...Rh1+! 31.Rc1 (or 31.Ka2 Nb3! (threat: Ra1#) 32.Nb1! =) Rxc1+! 32.Kxc1 Nd3+! 33.Kd2! Nxb2 34.Rc7+! Kf6 35.Rxa7 = <31.Ka2?!> For some reason, Kacheishvili decided not to play 31.b4! click for larger viewDIAGRAM  position after move 34 (34...Kf6). Technically here, the position is completely equal. Two grandmasters would probably soon have agreed to a draw, if not already. But when there is huge diffrence in strenght between two players, the stronger one will often play on and on, hoping that his weaker opponent will do a mistake along the way. After all, the rating is a number which illustrates the player's past performance. The higher the rating, the less mistakes were made, the less games were lost or even drawn, the less tactical calculation errors or strategical mistakes made. <36.f5!? e5! = / > This illustrates what was said in the previous paragraph. In a perfectly equal position, Kacheishvili cannot simply play the quieter and most evident moves if he is hoping for more than a draw (such as 36.b4 = or 36.Rac5 =). <37...Re7?!> The position contains some subtleties. This most evident move was to be avoided. Suggestion: 37...h6 38.b4 e4! = / . 

Jun1316   Sergash: <38.Ra6+! Kg5 39.f6! Rb7?!> Was Carlsen struggling to meet time control at move 40? 39...Re8 40.Rxa7 Kxf6 41.Rf2+! Ke6 (if 41...Kg6 42.Rfxf7! e4 43.a4 with the plan of opposing a rook on e7. Note that very bad would be 43...Rxg2?? 44.Rg7+ and Rxg2) 42.Ra6+! (if 42.Raxf7 Rxg2! =) Ke7! (if 42...Kd5 43.Rd2+! Ke4! 44.Rf6! ) 43.Re2 Kf8! = <40.Rxe5+ Kg6 41.Re7 Rb8!> White missed 41.Re8! with the idea of Re8g8+ and b2b4, supporting g2 from another spot. <42.Raxa7? Rxg2 = > Ouch! The first serious mistake of this game. Like Carlsen, Kacheishvili let go of his advantage that was so hard to obtain... 42.Re2 Ra8 43.b4 h5! 44.Kb3 <45...Rg4> Idea: outside the direct threat on White's 'a' pawn, if 46.a5 Ra4+ and Rxa5. Each side has at least a passed pawn. click for larger viewDIAGRAM. <51...Re6?? 52.b4 > Fatal mistake by Carlsen. Teachers repeat to put the rooks behind the passed pawns, yours and your opponent's. Here this would give: 51...Re1 52.Rg8+ Kf4 53.a6 Ra1+ 54.Kb5 h3 55.Rh8 Kg3 56.Rg8+ Kf3 57.Rh8 Kg2 58.Rg3+ Kf3 = etc. click for larger viewDIAGRAM. <54.a6?? Ra1+! = > Amazing, just when the grandmaster is winning the game at last... he loses any winning chance! He had to move the OTHER pawn: 54.b6! Rb1 55.Rh7! Rb2 56.b7 Kg6 57.Rd7 h3 58.a6 h2 59.Rd1 etc. 

Jun1316   Sergash: click for larger view<58...Rxa6?? 49.b7! > As was said above < wwall: 58...Rxa6? is the losing move. Black should be able to draw after 58...f2 59.Rg8+ Kh5 60.Rf8 Rxa6 61.Kc5 Ra5+ 62.Kc6 Ra2 63.b7 Rc2+ and 64...Rb2> 58...f2 59.Rg8+ (or 59.Rf8 Rxa6 60.b7 Rb6 61.Rxf2 Rxb7 =) Kh5 60.Rh8+ (or as mentioned by Wwall: 60.Rf8 Rxa6 and here I will say 61.Rxf2 Kxb6 =) Kg4 61.Rg8+ (or 61.Rf8 Rxa6 62.Ke3 Kxb6 63.Kxf2 =) Kh3 62.Rf8 Rxa6 63.Rxf2 Rxb6 64.Ke3 = only move. Black cannot win this position. <61...Kf4 62.Rf8+!> If 61...f2 62.Rg8+! Kf4 63.Rf8+ Kg4 64.Ke3 Kg3 65.Rf3+ Kg2 (or 65...Kg4 66.Kxf2 ) 66.Rxf2+ Kg3 67.Rf8! <63.Ke3 10.> Had Carlsen played a few more moves, it could have gone that way: A) 63...Kg2 64.Rg8+! Kh1 (or 64...Rf1 65.Kxf3 h3 66.Rh8 Ke1 67.Rxh3 etc.) 65.Kxf3 h3 66.Kg3 h2 67.Re8 Kg1 68.Re1# B) 63...f2 64.Ke2 h3 65.Rg8+ Kh2 66.Kxf2 Kh1 67.Kg3 h2 68.Re8 Kg1 69.Re1# C) 63...h3 64.Rg8+ Kh4 65.Kxf3 etc. 

Oct0519
  Messiah: 1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.Qc2 OO 5.a3 Bxc3+ 6.Qxc3 b6 7.Bg5 Bb7 8.e3 c5 <8...d6 is the most natural, and probably the best, with a lots of options for both sides. 8...h6 is also considerable.> 9.dxc5 bxc5 10.Ne2 Nc6
<A novelty over J L Chabanon vs Bacrot, 1995, which saw 10...h6. Probably 10...d5 makes the most sense, bombarding the center as soon as possible, and later playing against the uncastled king.> 11.Ng3
<11.OOO! Ne4 12.Bxd8 Nxc3 13.Nxc3 Rfxd8 14.Ne4 Ne5 15.Nxc5 is clearly better for White.> 11...Qa5 12.Bxf6
<The immediate queencapture was not bad, either.> 12...gxf6 13.Qxa5
<13.Ne4 was considerable, trying to improve the knight's placement.> 13...Nxa5 14.Nh5!?
<Strange move, but not entirely bad. Doing some expansion with 14.b4 was the way to go. However, White is just minimally better, thanks to the terrible placement of the bishop. Maybe 14...Nc6 15.Ne4 cxb4 16.Nxf6+ Kg7 17.Nxd7 was the best.> 14...Rfd8 15.OOO?
<I am unable to understand this decision once again, maybe Kacieshvili thought the rating difference will win the game automatically. Expanding with 15.b4 was in White's favour.> 15...Kf8 16.Nxf6 Ke7 17.Nh5?
<Why? The pawn is absolutely capturable.> 17...Ba6
<17...Rab8, preparing the devilish Be4 was considerable  Black would have been objectively better.> 18.Nf4 Bxc4 19.Bxc4 Nxc4 20.Rhe1 Rab8
<Or the immediate 20...d5 with ideas of Nd6.> 21.Re2 d5 22.Rc2 Na5 23.Kb1 c4 24.Ne2 Nb7?!
<24...Rd6, Rb3 or Nb3 were better: the central pawns will be very strong in later phases.> 25.e4 Nc5 26.exd5 Rxd5 27.Nc3 Rg5
<27...Rd3 is more natural for me, but it does not change the equality.> 28.f4 Rh5 29.Rd4 Rxh2 30.Rxc4 Nd3 31.Ka2 Rb7 32.Ne4 Rd7 33.Nc5 <32.Rd4 Ne1 33.Re2 Rxg2 34.Rxe1 Rbxb2+ 35.Ka1 Rb3 36.Nd5+ Kf8 37.Ne3 Rxa3+ 38.Kb1 Rb3+ 39.Kc1 Rc3+ 40.Kd1 Ra2 is wild, but I sense equality.> 33...Nxc5 34.Rxc5 Kf6 35.Ra5 Rb7
<35...h5 is an alternative, the position remains equal.> 36.f5 e5 37.Re2 Re7 38.Ra6+ Kg5 39.f6 Rb7?!
<Not understanding the position. The obvious 39...Rd7, with its eventual transfer to d2, was the way to hold an absolute equality  there is no need to pressure b2 and b3 at the moment, White's king does not want to leave a2 yet.> 

Oct0519
  Messiah: 40.Rxe5+ Kg6 41.Re7?!
<41.Re8 was psychologically better, asking Black to mess up something.> 41...Rb8 42.Raxa7?!
<Keeping the position complicated was psychologically better once again: 42.Re2 Ra8 43.b3 h5 44.a4, and although this is not necessarily winning, Black has a very wide room to lose the resulting endgames.> 42...Rxg2 43.Rab7 Rxb7 44.Rxb7 h5 45.a4 Rg4 46.Ka3 h4 47.Rb8 Rg3+ 48.b3 Kxf6 49.Rh8 Kg5 50.a5 Re3 51.Ka4 Re6?? <Completely forgetting a well understood and ancient rule: put your rooks behind the pawns! Now Black is lost. Draws were:> <<51...Re2 (or e1) 52.b4 f5 53.b5 f4,>> <<51...Kg4 52.b4 h3 53.b5 Re4+,>> <<and probably the easiest to see is 51.... f5 52.b4 Re1 (or maybe e2) 53.b5 f4, that is essentially the same variation as the first one.>> 52.b4 f5 53.b5 Re1 54.a6??
<A very serious blunder for a player rated 2500+, now the position is very drawish. After 54.b6 Rb1 55.Rh7 Ra1+ 56.Kb5 Rb1+ 57.Kc6 Rc1+ 58.Kb7 one of White's pawns will be promoted, significantly earlier than the f or hpawns.> 54...Ra1+ 55.Kb4 f4 56.Kc5 f3 57.b6 Ra5+ 58.Kd4 Rxa6?? <The losing move, a horrific miscalculation. 58...f2 was required:> <<59.b7?? f1=Q 60.b8=Q Qd1+ wins for Black.>> <<59.Rf8 is the only move (besides 59.Rg8+), and now Black can choose from several draws:>> <<<59...f1=Q (or =R) 60.Rxf1 Rxa6 is a wellknown draw,>>> <<<59...Ra4+ 60.Ke3 Rxa6 61.b7 Rb6 62.b8=Q Rxb8 is also dead equal,>>> <<<the immediate 59...Rxa6 60.b7 f1=Q 61.Rxf1 is also a draw,>>> <<<59.... Rf5 60.Rxf5+ Kxf5 61.b7 f1=Q 62.b8=Q is a theoretical draw once again, but why would anyone go for this line, especially in time trouble?>>> 59.b7 Rb6
<Now 59...f2 is not good anymore, as 60.b8=Q f1=Q 61.Qe5+ Qf5 62.Rh5+ wins.> 60.b8=Q Rxb8 61.Rxb8 Kf4 62.Rf8+ Kg3 63.Ke3
<Whatever Black does, the fpawn will be captured, and the hpawn is too slow to get promoted. White checkmates in every variations.> 10 

Feb1021
  Knighthawkmiller: Interesting ending in that if black advances the R pawn 63...h3 and white takes the pawn with check 64.Rxf+ is a draw. (but 64.Rg8+ wins).
=0.00 64...Kg2 65.Rf2+ Kg3 66.Rf5 Kg2 67.Rg5+ Kf1 68.Rh5 Kg2 69.Ke2 h2 70.Kd1 h1=Q+ 71.Rxh1 Kxh1 



