Kasparov - Karpov III, Game # 22

This is a game I started to analyze right after it was played. Thanks to some (maybe inaccurate) analysis that was published in one newspaper [chess] column, my initial opinion of this game was not as high as it probably should have been. (Beware other opinions, even those of titled players!  ALWAYS  think for yourself!!)

I found one analysis in a spiral-bound notebook. (Some of it was done shortly after this game was played. The bulk of it was added in 1991.) I incorporated some of this analysis and a few of my comments into my annotations of this fantastic encounter.

Many of you have written me to ask my opinion of this game. I had renewed interest in this clash when a GM wrote me about this game early in 2003. He stated he thought that this fantastic game might be the finest game Kasparov ever played ... and asked me to take a second look at it.  (My analysis took several months.)  

Unfortunately this game has been overshadowed by many others that Kasparov has played. This game is not in many of the more famous [book] collections of "Best Games," ...  all the more reason that I (and others!) should look at it carefully. Most of Black's moves are  VERY  plausible ... it is not entirely clear exactly where Black went wrong in this fascinating and very difficult game. (I have chosen to avoid knee-jerk question marks, here.)

Many mistakes have been made in the analysis of this game. Hopefully there are few errors in my work here. The reasons for this are twofold:  #1.) I took a great deal of time to perfect this analysis;  # 2.) All of my analysis has been extensively computer checked with the best chess programs. Much of my analysis is new - I copied it from no one.
(Where I did quote a line from another source, this is  clearly  indicated!!) 

For an explanation of the symbols that I use, please click  here.

This is mostly a text-based page, with only one diagram. Therefore, you will probably need a chess board. 

GM Garry Kasparov (2750) - GM Anatoly Karpov (2730) 
FIDE World Championship, (Return Match) 
London, ENG; ---> Leningrad, RUS (Game #22); 1986

[A.J. Goldsby I]

     The CB medal for this game. (kvsk-iii_gm22_med.gif, 02 KB)


The famous twenty-second (22) match game of the Kasparov-Karpov 
WCS Match. (This is their third match for the title.) 
It is a brilliancy that has few equals. The audience that witnessed the 
finale gave both players a standing ovation.

This game was picked as the best of the year by the Informant, 
the (USSR) magazine, '64' ... ... ... 
and many other publications as well. 

A GM, who is also a former U.S. Champion, e-mailed me to opine that this 
might be the finest game that Kasparov ever played. (Close study suggests 
it is a great game and a genuine masterpiece of the chess-board.) 


For my annotations here I have relied mainly on three books.
(A book of brilliant games by Dankin, {Danish? - mostly symbols} a book 
on great games, {chess brilliancy}; '250 historic games'  by Y. Damsky; 
 and the 1986 issue of the 'INFORMANT.')

I have been over this game literally dozens and dozens of times. 
It is incredibly brilliant. Most - or all - of Black's moves are very plausible. 
(The annotations here were close to 15 years in the making.)



1.d4 Nf6;  2.c4 e6;  {Diagram?} 
Black threatens a possible Nimzo-Indian.

3.Nf3,  {Diagram?} 
A slight variance in move order. 
(Avoiding the Nimzo.) 

     [ The more conventional move order would have been: 
        3.Nc3 d54.Bg5 Be75.e3 0-06.Nf3 h67.Bxf6! Bxf6; 
        8.Rc1,  "+/="  {Diagram?}  transposing back to the game. ]  


Black now pushes his QP ...  and transposes back into the main lines 
of a Queen's Gambit.
3...d5;  4.Nc3 Be7;  5.Bg5,  {Diagram?} 
Transposing to the main line of the QGD, but this move order 
 encourages Black to try the  'in-between move'  of ...h6. 

     [ Another try is:  5.e3!? ]  


5...h6!?;  {Diagram?}  
This is considered to be one of Black's most reliable options here. 
It is also a favorite of GM A. Karpov. 

   (Some books call this  ...  "The Petrosian Variation.")    

     [  More often seen is the following move order: 
        5...c66.e3 Nbd7; 7.Qc2!?{Diagram?} 
        The most modern try. 

         (The continuation of:  7.Rc1 0-0;  8.Bd3 dxc4;  9.Bxc4, {Diagram?}   
          leads to the variations most considered to be the time-honored    
          main lines. {See MCO or NCO.}   

        7...0-08.Rd1 a69.a3,  "+/="  {Diagram?}  
        White has a very small edge in this particular position. 

        S. Ivanov - A. LugovoiSt. Petersburg Champ. 
        RUS/1997.  (1-0, 64 moves.)  

        An interesting historic note - this position actually first occurred in 
        the game, (by transposition): R. Reti - R. Spielmann
        Trecianske Teplice, 1926. (!)  ]  


6.Bxf6!,  {Diagram?}  
The best move for White to try and keep the initiative. 

     [  The move  6.Bh4!?{Diagram?}  could lead to the  Lasker Variation

        (Black also might grab the initiative with an eventual ...g5.)  ]  


The next 5-6 moves feature pretty much standard development. 
6...Bxf6;  7.e3 0-0;  8.Rc1 c6;  9.Bd3 Nd7;  10.0-0 dxc4; 
11.Bxc4 e5!?;  {Diagram?}
A solid move. 

(A badly needed pawn break designed to give Black some 
 space to maneuver.) 

But! ...   in game twelve, Karpov had used the other pawn break 
of ...c5 instead. 

      [  Black could also play:  11...c512.Qe2 a613.Rfd1 cxd4; 
         14.Nxd4 Qe7!15.Ne4 Be516.Nf3 Bb817.Qd2, "+/="  {Diag?}  
and White has retained a very small edge, but the game was 
         drawn in 34 more moves. 

         GM G. Kasparov - GM A. Karpov; 12th Match game, 
         FIDE WCS/London/(ENG)/1986. ]  


12.h3,  (Maybe - '!')  {Diagram?} 
Kasparov plays to maintain as much play in this position as he possibly can.

     [  The continuation of:  12.Ne4 exd413.Nxf6+ Qxf6;  
"+/="  {Diagram?} 
         has been used by many GM's as a quick drawing line.  

         I.e., GM J. Benjamin - GM L. Christiansen;  
         U.S. Championship, 1986. ]  


Black now exchanges in the center, and White takes with the pawn 
to keep a grip on the center. 
(Taking with the Knight could lead to a quick draw.) 
12...exd4;  13.exd4, ('!')  {Diagram?} 
 To maintain a bind. 

     [ 13.Nxd4 Ne5; "~" ]   


13...Nb6;  14.Bb3 Bf5!?;  {Diagram?} 
This is OK. Keene calls it a  'nuance.' 
 (In his book on this match.)

     [  The move  14...Re8!?{Diagram?}  
        could transpose to a game these two contestants 
        played in 1985. 

       [ See also MCO-14;  page # 398, and mainly column # 1. ]  ]  


15.Re1 a5!?;  "~"  (TN?)  {Diagram?} 
"A position well known to theory."  - Yakov Damsky
 (But this is the first game in this position that I could find 
  {in this variation} by searching several on-line databases.)

This position is approximately level. (But! ... extremely complex!!) 

     [ 15...Re8!? ]  


16.a3 Re8!?;  {Diagram?}  
According to two books I have, this is the main line here. 

As far as I can tell, there hasn't been a huge amount of experience 
at the Master level with these lines. 

     Apparently Karpov can be found on both sides of this position.
16...Qd717.Ne5 Bxe518.Rxe5 Rfe8!?19.Qe2 Rad8; 
20.Re1 Rxe521.Qxe5 a4!?22.Qc5! axb323.Re7! Qd6?; {D?} 
        This looks foolish.  

            (Much better would have been:  23...Qc8; 24.Qxb6 Rd7; 25.Rxd7,   
             25...Qxd7; 26.Qxb3, "+/=" 
             and although White is better, Black has chances to draw.)    

        24.Re8+! Rxe8;  {Diagram?}  
        This is forced.

            (24...Kh7??; 25.Qxf5+ Qg6[]; 26.Qxg6+ Kxg6;  27.Rxd8, "+/-")     

        25.Qxd6, "+/-"  {Diagram?}  
        White is winning.  

        GM A. Karpov - GM A. Belyavsky;  European Cup Champ, 1986.  ]  


17.Rxe8+ Qxe8;  18.Qd2, ('!')  {Diagram?}  
The best way for White to try and keep an advantage.

     [ 18.Bc2 Bxc219.Qxc2 a4; "=" ]  


18...Nd7!?;  {Diagram?} 
This could be less than best. 
(Damsky gives it a dubious.)
Indeed, experience has shown that Black will come closer to 
equalizing by getting this Knight to the d5-square ... and the sooner,
the better. 

Karpov only took about five minutes on this move and his demeanor 
clearly indicated he was very comfortable with his position up to this 
point in the game. (Possibly he was still in his pre-game preparation?) 
 {See the book by Keene & Goodman on this match.}   


     [  Black could play:  >/=  18...Qd7!?{Diagram?}  
        MCO stops here and calls this position pretty much even.

        H. Olafsson - van der Sterren;  Wijk aan Zee, 1987.

        [ See MCO-14; page # 398, column # 1, and also note # (d.). ] 

        19.Re1 Re820.Rxe8+ Qxe821.Qf4, "+/="  {Diagram?}  
         M. Gurevich - P. van der Sterren;  Baku/RUS/1986. 


        Maybe  18...a4!?; "~"  {Diagram?}  
        but this would transpose to many of the games that have 
        been played in this line.  {A.J.G.}  ]  


Now White maneuvers very skillfully and slowly increases his edge and 
demonstrates his slight advantage in space. 
(Many of these moves are not at all obvious.) 

19.Qf4! Bg6!?;  {Diagram?}  
Keene says that ...Be6 had to be better, but theory, (and most of the GM's 
who have bothered to really analyze this position); says otherwise.

"19...Be6; is a minor victory for the player of the White pieces." 
   - Garry Kasparov

     [  19...Be6!?; "~"  {Diagram?} 
Originally I stopped here and stated that the position is more 
         or less unclear. (I stand by that particular evaluation.)  

         The following piece of analysis is mostly by Kasparov & Karpov. 
20.Bxe6! Qxe621.Qc7! Qb3!?;  {Diagram?} 
         Other move achieve little.

           (21...Rb8; 22.Qxa5, "+/=")     

         22.Ne4! Qxb223.Re1! Nf824.Nxf6+ gxf625.Qf4!?, "+/="  {D?}  
         White has both an advantage and a nice attack against the greatly 
         weakened position of the Black King.  

           (25.Re7, "+/=")    ]  


20.h4!,  {Diagram?} 
Kasparov calls this a valuable multi-purpose move which is 
handy for both the attack and restraining Black's possibilities. 
(And his pieces.) 

     [ 20.Re1!? ]  


20...Qd8;  {Diagram?}  
Black's position still seems close to being OK.

With his next move, White stops Black from playing ...Qb6; 
and virtually provokes Black into (eventually) playing the pawn 
advance,  ...b5.
21.Na4! h5!?;  {Diagram?}  
Karpov said after the game that this was (more-or-less)  
positionally forced to stop White from playing g2-g4.  
(See the third variation just below.) 

Of course...h5 has the slight drawback of just minorly weakening
the Black King-side, however imperceptibly. 


      [  21...Rc8!?; "~"  {Diagram?}  
         The position is unclear.


         Or  21...Qc8?!22.Ne5, "+/="  {Diagram?}  
White is clearly for choice here. 


        Or  21...Bh5!?22.g4, (!)  22...Bg623.h5, Bg6;  
"+/="  {Diagram?} 
with a clear advantage and an initiative for White.  ]   


22.Re1 b5;  {Diagram?}  
Black grabs some Q-side space here. 
(And he also kicks the White Knight - with a gain of time.) 

23.Nc3,  {Diagram?}  
This retreat is the correct play here, in this position.

     [  A horrible mistake is:  </=  23.Nc5? Nxc524.dxc5 a4!;  
        25.Ba2 Bxb2; "/+"  {Diagram?}  
        which is (very close to) winning for Black. ]   


23...Qb8;  24.Qe3!,  (Maybe - '!!')  {Diagram?}  
The strongest move, and given an exclam by  GM R. Keene.
(And others.)

     [ 24.Ne5!? ]  


24...b4;  25.Ne4 bxa3;  {Diagram?}  
Black tries to relieve some of the pressure.

It is not advisable for Black to try and win a pawn here.

     [  Much too risky is:  25...Bxe4!?; ('?!')  26.Qxe4 bxa327.Qxc6!!, 
        27...axb228.Qd5!, "+/-"  - Kasparov. ]  


26.Nxf6+ Nxf6;  27.bxa3 Nd5!;  {Diagram?}  
The best move, and this is awarded an exclam by many
different authors here.

  '!'  -  I. Damsky.   

     [  27...Ng4!?28.Qc3 Qd629.Ba4!, "+/="  - GM A. Karpov]  


White's next move is virtually forced. We now reach a position where 
White's Knight is vastly superior to the Black Bishop. 
28.Bxd5 cxd5;  29.Ne5 Qd8!?;  {Diagram?}  
Perhaps better than ...Qd6.  

     [ 29...Qd6!? ]   


30.Qf3!,  {Diagram?}  
Kasparov was to later remark that one might have expected play on
the dark squares here with Qf4. (Or Rc1.)  But analysis, intuition, ... 
or maybe just pure genius ... must have led him to find that this square 
was somehow superior. 

--->  One funny story - that I did not know - was emailed to me by a player who 
was present when this match was played. Apparently a very important and 
respected arbiter nearly stopped the proceedings at this point in the game. 
"Surely the demonstration board is incorrect? How could Garry not have 
 played the very strong move, Qf4?" he asked. 

     [  Also good was:  30.Qf4!?, "+/=" {Diagram?}  
         with a small but solid advantage for White in this position.  ]  


30...Ra6;  {Diagram?}  
"Probably the only move."  - GM A. Khalifman.  
 (Black wishes to avoid the doubling of the pawns in front of his King.)

     [ 30...Rc8!?; 31.Nxg6, "" ]  


Garry could (now) play Nxg6 ... with a clear advantage here. Instead 
he chooses to slowly ratchet up the pressure in this position.  
31.Rc1! Kh7!?;  {Diagram?}  
Damsky questions this seemingly obvious move, and gives instead
{the seemingly more risky} ...Qxh4. This is based on a lengthy analysis 
that was published in a Russian magazine, but may contain several
errors. (Apparently Kasparov was of the same opinion.)

It seems no matter what move Black plays in this position, White gets a very 
clear and convincing advantage from this point in this exciting contest. 

     [  31...Rf6!?32.Qc3, "+/=" {Diagram?} 

        Or  31...Qxh4!?32.Qxd5, "+/=" {Diagram?} 
        (Maybe  - "+/")  ] 


32.Qh3! Rb6;  33.Rc8 Qd6;  34.Qg3! a4!?; {Diagram?}  
A reasonable-looking move that is branded as dubious by Damsky.  

     [  The continuation of:  34...Rb1+!?35.Kh2 Qa636.Re8! Qf1!?;  
         37.Qf3 a4?;  {Diagram?}  This is bad. 

          (The move, 37...f6;  had to be played in this position.)   
           {After ...f6; White wins with Rh8+!!} 

         38.Nxf7! Bxf739.Qxf7 Qg1+40.Kg3 Rb3+;  
         41.f3, "+/-"  {Diagram?}  
         wins for White, as the e1-square is now guarded. 
         - G. Kasparov]  


35.Ra8!,  {Diagram?}  
"An unpleasant surprise for Black in this position." 
 - GM Garry Kasparov.
 (The a-pawn cannot be defended.) 

     [ 35.Kh2!? ]  


35...Qe6;  {Diagram?}  
Karpov felt this was forced.  

     [  Variation # 1.) 
        A cute tactic is:  35...Ra6; ('?')  36.Nxf7! Bxf7; {Diagram?} 
        This is forced. 

          (Not  >/=  36...Qxg3??;  37.Rh8#)   

        37.Qd3+;  ("+/-")  {Diagram?}  which wins for White. 


        Variation # 2.) 
        Also bad for Black is:  35...Rb3?!36.Rh8+! Kxh837.Nxf7+ Bxf7;  
        38.Qxd6, ''  {Diagram?}  and most programs consider White's 
        position to be winning ("+/-")  here.  ]  


36.Rxa4 Qf5;  37.Ra7!!,  {Diagram?}  
For well over ten years, the best computer programs all pick (the move) 
Ra8 here - in this position. {A.J.G.} 
 (Many good players choose this move as well.) 

"Which is better, Ra7 or Ra8? I judged that the Rook would be capable of 
  more here. In particular, it secures the d7-square for the White Knight, 
  the appearance of which at f8 will be fatal for the Black King."  
  - GM G. Kasparov.

Most annotators give this move only one exclam ... but given the extreme 
depth of White's plan ... and the fact that most GM's who were present at 
this contest did not have a clue as to what was happening here ... I am 
convinced it fully deserves more.  {A.J.G.}  

     [ 37.Ra8!?, "+/=" ]  


37...Rb1+;  38.Kh2 Rc1;  {Diagram?}  
Karpov played this move very quickly, indicating he probably 
thought it was forced.  

     [  38...Rb2!?39.Nf3!,  {Diagram?}  
        Garry says this is best.  

          (Also good for White is: 39.Nxg6 fxg6;  40.Qe5, "")      

        39...f640.Qc7, "+/="  {Diagram?}  (Maybe "+/") 
        - GM G. Kasparov]   


39.Rb7!,  {Diagram?}  
This prevents ...Qb1. Garry could get a HUGE edge in the ending by 
first exchanging on g6, and then by playing Qe5, but he obviously 
wants more from this position.

     [  39.Nxg6 fxg640.Qe5, ""  {Diagram?}  

        </=  39.a4? Qb1; "<=>"  {Diagram?}  ]  


39...Rc2;  {Diagram?}  
Black doubles up on the pawn on the f2-square.  
(Black has few playable options.)

      [ </=  39...Rf1?40.Nf3 f641.Qc7, '' ]  


40.f3,  {Diagram?}  
The most precise, and if you wanted to - you could also give 
 this move an exclamation point.

     [ 40.Nf3!? ]  


40...Rd2;  (Maybe forced.)  {See the diagram just below.}  
The game is very tense. The last few moves were played rather 
quickly, as both players were running short of time. 
This was the adjourned position. 

ALL (!!!) the GM's who were watching this game believed Black stood 
very well here. Even Kasparov's own team was rather glum, fully feeling 
that Black stood very well at the break.  

   The sealed position - most GM's missed what was to happen next.  (kaspvskarp-iii_gm22.gif, 58 KB)

(The actual game position after 40...Rd2.) 


41.Nd7!!,  (Maybe - '!!!/!!!!')  {Diagram?}  
The best move ... and the one that Kasparov sealed here ... meaning  
that it had to be discovered while he was thinking at the board, and is 
not the product of adjournment analysis.  

One of the most shocking and most brilliant of all sealed moves in the 
history of the game. (of chess) 

ALL the annotators are universal in their praise of this move ... and the 
fact that this move fully deserves two exclamation points.
(Even a jaded author like Keene gives this move two exclams!)

     [  After the continuation of: 41.Rb4 f642.Nxg6 Qxg6;  
         43.Qxg6+ Kxg6;  {Diagram?}  
         Soviet analysis has shown that Black could probably 
         draw this ending. 
         (With perfect play, I hasten to add.) 


        The move of:  41.Nxg6!?{Diagram?}  
         will probably transpose into the line that is given above. ]  


41...Rxd4;  42.Nf8+ Kh6;  {Diagram?}  
This is forced, after ...Kb8??; Rb8 wins instantly. 

Another completely unexpected shot.
All the Masters who were present were busy analyzing Rook to b6.
(Or they were examining the consequences of Nxg6.)

White prevents an exchange of Queens on f4 and tightens the noose 
 around the Black King.

     [ 43.Rb6, "+/=" ]  


Karpov thought for a bit, but apparently could not find any way to 
save his position.
43...Rc4!?;  {Diagram?}  
GM Raymond Keene calls this the line of least resistance.
{He labels it as doubtful.}
(But Black has no good moves here.)

     [  Variation # 1.)  
        Kasparov  provides the following entertaining line here: 
        43...Rxb444.axb4 d445.b5 d3;  {Diagram?}  
        There is little else that Black can do here, his Queen cannot 
        wander from its current post. 
        (A check on f4 is curtains.)

           ( </= 45...Qxb5??;  46.Qf4+, ("+/-") )     

        46.b6 d247.b7 d1Q;  {Diagram?}  
        Usually the player who promotes first wins handily ... but not here!! 

        48.b8Q Qc1;  {Diagram?}  
        Black had to prevent a check on the c1-h6 diagonal. 

        49.Nxg6 Qxg650.Qh8+ Qh751.Qgxg7#. 


        Variation # 2.)  
        Plain silly is:  43...Rd1?44.Rb8 Bh745.Qg5+ Qxg5;   
        46.hxg5+ Kxg547.Nxh7+,  ("+/-")  {Diagram?}  
        and White wins a piece. 


        Variation # 3.)   
        A long - but relatively simple win - is to be found in the line that follows: 
        43...Rd344.Rb8,  {Diagram?}  
        The most direct.  

          ( 44.Qe1! "+/=" - Keene. )     

        44...Bh7;  {Diagram?}  
        This is forced.  

           (44...Rxa3??;  45.Nxg6 fxg6?; 46.Rh8#)     

        45.Qg5+,  {Diagram?}  
        With this simple tactic, White wins a piece.  


        45...Qxg546.hxg5+ Kxg547.Nxh7+ Kh648.Nf8 Rxa3; 
        49.Rb7 f650.Rd7 Rd351.f4 Rd152.Ne6 Kg653.Nxg7 h4!?;  
        The only practical chance. 

        54.Kh3 f555.Kxh4, "+/-"  {Diagram?} 
and White obviously has a won game.  ]  


44.Rxc4 dxc4;  45.Qd6! c3;  46.Qd4,  {Diagram?}  

Black Resigns.

The position is completely hopeless. 


Several  (most!)  of the judges for the panel of the Informant, scored 
this game as nearly perfect. (!!!) 

Living legends of the game ...   
- such as  Botvinnik, Tal, & Smyslov  - also  greatly  praised this game.

Najdorf,  who was normally a very harsh critic of most of the Russian 
players, (and the way they often play chess);  called this game: 
 "exceptional and exquisite." 

Garry Kasparov, perhaps understandably exuberant about this game, said: 
"In time the score of this (return) match may be forgotten, and perhaps even 
 the name winner; but the move  Ne5-d7  ...  and everything associated with it  ... 
 will remain in chess forever." 
 (Heady words to be sure, but possibly he is right.)


   Copyright (c) A.J. Goldsby I.  Copyright (c) A.J.G;  2003 & 2004.   


     [  After the continuation of:  46.Qd4 Bh7;  {Diagram?}  
        This is forced. 

         (46...c2??; 47.Qe3+,  and Black will be mated.)    

         47.Qxc3 g5{Diagram?}  
This looks close to being forced here. 

           (Slightly worse for Black is the line:    
             </=  47...Bg8; 48.Qe3+ g5; 49.hxg5+ Qxg5?!; 50.Qxg5+,      
             50...Kxg5; 51.a4, "+/"  (Maybe  "+/-") 
Iakov Damsky.

         48.a4!,  {Diagram?}  A brilliant coup.

         ( Or 48.Qe3, '' )    

         48...gxh4;  {Diagram?}  
         Black has few good moves in this position. (zugzwang)  

         49.Qe3+ Qg550.Qxg5+ Kxg551.Nxh7+ ("+/-")  {Diagram?}  
         and White's position is one that any amateur could win. ]  


   (Code Initially)  Generated with  ChessBase 8.0   

  1 - 0  


    Copyright () A.J. Goldsby, 2004.  All rights reserved.  

May 22, 2003:  Tuesday I received an e-mail from a GM. He gave me his name, but did not give me permission to use it. He told me he went over this game. While he confessed that his better playing days might be long gone, he said he enjoyed my analysis. He also told me this was some of the best chess work/analysis that he had ever seen ... anytime, anywhere. He also wished me well. 

There is no site map, but you can click  here.  

 Click  here  to go to - or return to - my page of:
ANNOTATED GAMES, (Angel-Fire 2) Page #2

Click  here  to return to the main  "Recent GM Games"  
(home/main)  page. 

(To contact me concerning this analysis, click here.)

This page was last updated on 05/15/06 .

 Copyright ()  A.J. Goldsby, 2006. 
  All rights reserved.