J. Capablanca - S. Tartakower;
NY, 1924

(This is the finest analysis of this game you are ever likely to see on the net.
Or ... anywhere else, for that matter!!!)

I annotated this game (primarily) for my web page on the, "Best (and Most Amazing) [Chess] Games - Ever Played."

Click HERE to see this game on a beautiful replay board. {Search for this game on a popular chess website.}

Explanation of symbols: The symbols I use are the standard ChessBase/Informant ones.
(Except they are more-or-less turned on their sides, and with a few modifications.)

A "plus-slash-minus" ("+/-"), means that, "White is winning."
A "plus-over-a line" ("+/"), means that White is much better.
A "plus-over-an equal sign" ("+/="), means that "White is just slightly better."
An equal sign ("=") means the game is level or the play is roughly balanced.
An "unclear symbol" ("~") means the play is unclear, but should be roughly close to equal.
A "minus-slash-plus" ("-/+") means that, "Black is winning."
A "plus-under-a-line" ("/+"), means that "Black is much better."
A "plus-under-an equal sign" ("=/+"), means that "Black is just slightly better."
The symbol "<=>" means, "with good counterplay."
The symbol "~" means, "good compensation" (i.e. active play), for the material invested.
The symbol "--->" means, with a strong attack.
The symbol "/\" means, with a strong initiative.
(This is not quite the same thing as an attack, but means basically the ability to greatly dictate the course of play.)

I also use the mark, (appellation) '!?' a lot. The normal interpretation of this mark (after a move) is interesting, but perhaps associated with a great deal of risk. I do not agree with this interpretation and prefer just interesting. I also use it to show a divergent path in the analysis, or a departure from normal opening theory. This does NOT mean the move is good or bad, just interesting ... and there are possibly MANY different alternatives at this particular juncture of the game!!!

GM Jose R. Capablanca (2740) - GM Savielly G. Tartakower (2650)
[A80]
The New York International Chess Tournament
New York, NY (Rd. # 6), 1924

[A.J. Goldsby I]

( The ratings above are actually very conservative estimates. They have been adjusted to reflect modern-day standards and the constant inflation of the ratings since ELO first introduced his system. By comparison, the best players of 2002 are nearly 2800, (or over!); and the average "Top 50" GM is usually in the rating range of 2675 ... give or take 20-30 points either way.)

A great game! In fact one of the perfect model games demonstrating the value of an active King in the ending of King and Rooks. (Plus pawns, for both sides.) {A.J.G.}

After DECADES of reflection ... (I first saw this game when I was 8 years old!); I must say this is one of Capa's GREATEST GAMES!! (An incredible statement in itself!)

As a teacher, I'd have to say this is also one of, "THE TEN MOST BEAUTIFUL AND INSTRUCTIVE END-GAMES EVER PLAYED IN THE HISTORY OF CHESS!!!"
(Especially in the category of R+P endings!! Maybe even the finest Rook-and-Pawn ending ever played!!!)

Irving Chernev says of this game:
"If it is indeed true that Capablanca studied more than a thousand rook-and-pawn endings to refine his incomparable technique, then in the ending of this game he offers us the quintessence of his acquired and inborn mastery."

"This ending, which may very well be the finest and the best of it's kind, is one to delight the connoisseur, and one to grant the student a priceless lesson in the art of finishing a game with elegance and accuracy." - The (late) great, IRVING CHERNEV.
( From the introduction to this game, in his book, "The Golden Dozen." - - - 'The 12 Greatest Chess Players Of All Time.' - - - {Plus selected, annotated games.} Copyright 1976, Oxford University Press. Library of Congress # 75-39504. )

(See also Chernev's other book, "The Most Instructive Games of Chess Ever Played," by I. Chernev. Copyright, 1965. Game # 1.)

[ I have also gone back and referenced two more books, concerning this wonderful game:
# 1.) "The 100 Best," by GM Andrew Soltis.
# 2.) {The Mammoth Book of} "The World's Greatest Chess Games."
By GM John Nunn, GM John Emms, and FM Graham Burgess. ].

(My annotations are based on dozens of books. (And magazines.) ... ... ... But the best is easily the book, "The Golden Dozen," by I. Chernev.)

The comments in the double brackets (<< blah-blah-blah >>) are from the short annotated version of this game, which I plan on trying to have published.

( I have examined this game dozens - if not hundreds of times - over the years, and my notes will reflect this.
- LIFE-Master A.J. Goldsby I )

Many different sources give this game as starting out with the 'French' move order.
( I think database technicians call this, 'normalizing a move order.' )

This has been a common idea for over one hundred years with Dutch players, as Black can
then avoid the Staunton Gambit.

[ Chernev (& Soltis) gives the move order: 1...f5; 2.Nf3, The most restrained.

(The line, 2.e4!? fxe4; 3.f3 exf3; 4.Nxf3, leads to the somewhat dangerous Staunton Gambit.
This is something many Masters prefer to avoid!
Another way to play this opening is: (current theory) 2.c4 e6; 3.Nc3 Nf6; 4.Nf3 Bb4;
5.Qb3 Qe7; 6.g3 b6; 7.Bg2 Bb7; 8.0-0 Bxc3; 9.Qxc3 0-0; 10.Bg5, "+/="

   Or 2. c4, Nf6; 3. g3, e6; 4. Bg2, Be7; 5. Nf3, 0-0; 6. 0-0, d6; 7. Nc3, Qe8; ('!?')
   8. Re1, Qg6; 9. e4!, fxe4; 10. Nxe4, Nxe4; 11. Rxe4, Nc6; (Taking the Rook loses
   to Nh4, trapping the Queen.) 12. Re1, e5!?; The end of the column.
     (Maybe better was 12...Bf6.)   13. dxe5, Bg4; 14. h3, Bxf3; 15. Bxf3, Nxe5;
   GM A. Yermolinsky - Zelensky; USA, 1997.
   16. Bxb7, "+/" [See MCO-14; pg. 487-488, column # 17, note # (o.). ]. )

2...e6; 3.c4 Nf6; 4.Bg5, transposing back to the game/move order given here.
Since so many books give so many conflicting move orders for this game, and since it really does
not matter! ... (nothing is omitted or missed!); I have stayed with the move order given here.
{ALSO ... this move order is the way this game is listed in most electronic databases!}

BUT ... if I had to make a choice ... I would place my bets on Chernev!!
(He was extremely meticulous in his research.)

This is also a problem that repeats itself in nearly ALL electronic databases, the game scores
are often the wrong move order, ... or simply incorrect!! ].

2. Nf3,
Capa shows that he prefers a QP-type opening. This move also discourages any attempt at
a freeing central advance, [...pawn (at) e7 to e5;] anytime soon by Black.

The continuation: 2.c4!? f5; 3.Nc3 Nf6; 4.a3!? Be7; 5.e3 0-0; 6.Bd3 d5!?; 7.Nf3!? c6;
8.0-0 Ne4; 9.Qc2 Bd6; 10.b3 Nd7; 11.Bb2, "+/=" was the game,
GM Geza Maroczy - GM Savielly Tartakower;   Teplitz-Schonau; 1922. ( 0-1 in 35 moves.)  [replay]
This led to a game that was super-brilliant, won the brilliancy prize, and was perhaps the best
game of Tartakower's career. I mention this game only to make a point: It clearly demonstrates
that GM Savielly Tartakower regularly used the Dutch Defense ... and was very effective
with this particular opening! ].

2...f5; 3. c4;
The standard, "classical" way of playing this position. (White grabs some space and exerts
strong pressure on the center before playing Nc3.)

[ A more modern treatment of this opening would be: 3.g3 Nf6; 4.Bg2 d5!?; 5.c4 c6;
6.0-0 Bd6; 7.b3!? Qe7; 8.Bb2 0-0; 9.Nbd2, "=" with very complex play, and an
interesting middle-game to come. ].

3...Nf6; 4. Bg5!?;
Again, a simple - but effective! - classical deployment in the opening phase by White.
(White wishes to play e3, and develop his King-side, but does not want to close in his
QB behind his pawn structure.)

Capablanca may have been the ULTIMATE pure, classical player. (He generally almost
always seem to prefer a classical method of development in the opening phase of the game.)

[ Could White have tried 4.g3, in this position? (This is the more modern method for
handling this opening.) ].

[ Alekhine suggested: 4...Bb4+; 5.Nbd2 Nc6; 6.e3 0-0; with the idea of ...d6; and ...e5.
But this is not clearly better than the text. ( 7.a3, "+/=" ) ].

5. Nc3 0-0; 6. e3 b6;
Black plans to fianchetto his QB, a common idea in these lines of the Classical Dutch.
(This hyper-modern deployment of the QB is seen in even the most modern theoretical
treatments of the Dutch Defense, even today.)

7. Bd3 Bb7;
"Black's set-up is quite reasonable." - FM G. Burgess.
(He goes on to point out the pros and cons for both sides.)

Suffice it to say that both sides have arrayed their armies in an apparently normal and
acceptable manner. (Center, piece development, etc.)

8...Qe8!?; (Hmmm.)
(Thematic.) Black gets ready to transfer his Queen to the Kingside. The Queen will be used to
initiate an attack on White's King, if possible. This is also a common theme in the Dutch Defense.

According to Chernev, this was GM Tartakower's improvement on an earlier round of the
same tournament. (A TN?)

[ 8...Ne4; 9.Bxe7 Qxe7; 10.Bxe4 fxe4; 11.Nd2 Qh4; 12.Ndxe4, cost Black a pawn.
    GM F. Marshall - GM S. Tartakower; New York, 1924. (12...Bxe4; 13.g3 Rf6;
    14.f4, Seemingly forced.   (14.gxh4?? Rg6+; 15.Qg4 Rxg4#)   14...Qh3; 15.Nxe4 Rh6;
    16.Qe2, and White safely keeps the pawn.).

Or 8...Nc6!?; 9.Rc1 h6; 10.Bh4 Ne4; ("=") ... "with equality." - GM A. Soltis.
GM J. Nogueiras - GM A. Yusupov; Rotterdam, 1989. ].

9. Qe2!, "+/="
This threatens e4. (And causes, "an immediate change of plans," according to Chernev.)
The advance, e4 would give White the edge because of a huge, "spatial edge in the center,"
says GM A. Soltis.

'!' - Irving Chernev. '!' - GM A. Soltis.
'!' - FM Graham Burgess. (& GM/Dr. John Nunn.)

[ If 9.a3!?, then 9...Qh5; "~" is much better for Black than the actual game. Now play could
proceed 10.b4!? Ng4!?; 11.Bf4, "+/=" (Not 11.Bxe7?? Bxf3; "-/+" This is a typical
Dutch tactic, or type of attack you will see in this particular opening.) ].

9...Ne4; (Maybe - '!') {The position is close to equal.}
Tartakower certainly knows the value of a blockade.

[ 9...Qh5; 10.e4!?, "=" with advantage to White, according to Chernev.
(The position might actually be closer to equal.) ].

[ 10...Qxe7?; 11.Bxe4 fxe4; 12.Nd2 d5; "+/" gives Black a horrible pawn structure. ].

(Now White must capture the Black Knight on c3 - his Q is attacked.)
11. bxc3 Qxe7;
Black threatens an invasion on the Queenside on the a3-square. (White prevents it.)

"To make up for his Pawns being doubled on the Bishop-file, White has the (half) open
Knight-file for his Rooks." - I. Chernev.

12. a4!,
Prevents an invasion of the Black Queen on a3, and any thoughts of Q-side expansion by Black.
It also prepares a possible advance of a4-a5. This would swap off an isolated QRP for White
and perhaps create more avenues of attack. (The move also gains a little space on the Q-side.).

<< A multi-purpose move. One main point is to keep the Black Queen off the a3-square. >>

12...Bxf3; ('!?') {Diagram ?}
White was planning to play Rfb1 and c5 un-doubling his pawns, so Black exchanges
off his Bishop. (...Nc6; first hems in his Bishop.)

While this may or may not be true, it is certainly easier to play White in this position than
it is to play Black!

And to the best of my knowledge, no other annotator may have pointed out that
Tartakower may have been trying to create a situation where he could use his Knight
to exploit White's doubled QBP's. (This is {maybe} the way that Nimzovich himself
may have wanted to handle this position.) This certainly seems likely, considering the
further course of the game.

This is one of the main turning points of the game and it was a very critical decision.
It also sets up a major imbalance in the position, and it is an imbalance that will play a
major role in this game.

<< Black sets up a major imbalance with this move. Perhaps he was trying to exploit the
doubled White QBP's? >>

[ 12...Nc6; {Diagram.} This blocks Black's own Bishop, and could be why Tartakower
     swaps off his Bishop first. 13.Rfb1!,   (13.e4!?)   13...Na5;    (13...h6?!; 14.c5, "+/=").
     14.c5, "/\"   This clearly favors White, according to Chernev and several computer programs.
     14... bxc5?; 15.Rb5, "~" (Maybe "+/=")

13. Qxf3,
"Now it happens - as usual in mobile pawn formations - that the Bishop is superior to the
Knight. The rest of the game is a very fine example of the utilization of such an advantage."
- GM A. Alekhine. (From the book of the tournament.).

I think what Alekhine means is that the ending in this game, (and White's subsequent
advantage); is a direct result of all the pressure White obtained in the middle-game.

The very good writer, IM J. Silman, says that such positions should always be measured in
terms of it's major balances and imbalances.

The most notable imbalances here are:
A.) White's superior minor piece;
B.) White's small - but clear - advantage in space;
and C.) White's slightly more {potentially} active and mobile pawn center.

Since many computers rate this game {right now, in this position}; as dead equal, perhaps the
most impressive thing about this game is the slow, but sure - and very careful and methodical,
- manner in which Capa manages to gradually increase his advantage to the point where it is a
real and a tangible influence!

[ There is no need to play: 13.gxf3?!, ('?') and ruin White's pawn structure. Now Black
should play: 13...e5!, "=/+" with a very small (but clear!) plus. ].

14. Rfb1!, (Maybe even - '!!')
A unique maneuver - and the very best move here, according to Chernev.

14...Rae8; Centralization.
Black also possibly thinks about preparing the advance of ...e5; in the next few moves.

[ 14...g5!?; - Alekhine. This looks (possibly) risky to me.
14...Na5; - Tartakower. 15.c5 bxc5; 16.Rb5 c4; "~" ].

15. Qh3!, (Maybe - '!!').
This prevents ...f4, because of the double hit on h7. (And White is preparing f2-f4, to
prevent Black from making a central advance and freeing his game.).

This is a maneuver several Masters failed to discover when I showed them this game.
( I have tested this game on hundreds of players and dozens of computers over the years.
I also have taught this game more times than I care to count.)

<< White hits f5, preventing both ...e5; and ...f4. He is also preparing f2-f4, clamping down
on the center. >>

15...Rf6; (Maybe - '!?')
Black looks for counterplay. (Attack the enemy king?)

(Black plans ...Rh6; perhaps in the hope of dissuading White from executing the maneuver
he actually played in the game.)

Several players have criticized this move, most notably Alekhine. Yet no one, to the best
of my knowledge, has ever provided a move (or a plan!) that is distinctly superior to this one.

Also, the computers show virtually no change in the evaluation of the position after
15...Rf6. This would mean there is absolutely nothing deficient about this move!

Additionally, no other author; (Soltis, Chernev, Nunn, Burgess; et al) finds any fault with this
move at all. Personally, I think too many authors engage in illogical WITCH HUNTING.
What I mean by this - is in a game such as this, where there ARE NO EASY ANSWERS -
too many people have gone looking for ONE MOVE. One move to blame all the troubles
of the game on. One move on which they can hang their ... 'scarlet letter' on. BUT .....
there are no simple solutions. The hard, cold fact is that there may be NO one move in this
game which is to blame for all of Black's problems in this particular crown jewel of a game.
The simple fact is that the entire PLAN (and set-up) that Black came up with in this game
was ... REFUTED!! (But only by the most beautiful and exquisite play, I hasten to add.)

AND ... another thing to consider. This is a game I have been testing computers on, ever
since the computers have attained reasonable playing strength. (One of the first boxes that
I tested this particular game on was a Fidelity Electronics machine rated around 1700. That
poor thing had not a clue and rarely found a good move, especially as concerns this game.)
Even today, in the year 2002, NO computer program can successfully duplicate all the
moves in this extra-ordinary game. What I am trying to say is that Capa's play is sheer
genius here. It is for us to admire and enjoy --- and emulate ... if we can. But to capture it,
define it, or to get a machine to reproduce it ... THIS SIMPLY IS NOT POSSIBLE!!
Just as no one can be sure where the well-springs of real genius like Mozart, Beethoven,
Rembrandt, Da Vinci, or Einstein comes from; we cannot be sure where the source of
Capa's true inspiration flows from ... or even where it is going! We can only know, when
we view the games of this incredible player, we are looking at true genius! So ENJOY!

[ A blunder for Black is: 15...f4??; 16.Bxh7+! Kf7; (16...Kh8; 17.Bg6+ Kg8; 18.Qh7#).
17.Qh5+ Kf6; (17...g6; 18.Qxg6#). 18.Qg6#; Not 15...e5??; 16.Bxf5, "+/"

16. f4!, {Excellent!}
White gains space and mechanically prevents the advance of Black's f-pawn. He also puts
a virtual LOCK on the key central squares. (The pawn break, ...e6-e5; is now practically
impossible for Black to achieve.).

16...Na5; (Maybe - '!?')
Black tries to block White's Q-side play. (The 'Nimzovichian' concept of a Knight Blockade.).
(Additionally another concept that MANY annotators fail to point out ... and MOST players
miss here, is that Tartakower was hesitant to play the move, ...d6. This is because he feared
Capa may have found a way to exploit the now unprotected Knight on c6, perhaps by pinning it.
Ideas of c5!? and Bb5, and/or Qf3.)

[ MANY annotators have suggested the move:
    16...a5!?; This causes a weakness which could be exploited by ... 17.g3 g6;   (17...g5!?).
    18.Qf1 Qg7; 19.Be2 g5; 20.Kh1 gxf4; 21.gxf4 Kh8; 22.Bf3 Ne7; 23.Rb5 Rg8;
    24.Qf2 Rh6; 25.c5 bxc5; 26.Rxc5 Ra8; 27.Rb1 c6; 28.Rb7 Nd5; 29.Bxd5 cxd5;
     (29...exd5; 30.Rxa5!?).   30.Rcb5, and Black's a-pawn is rather weak the coming end-game.

Black could also try 16...d6!?; but after 17.Qf3, Na5; we have transposed back to the game. ].

Now that White's Queen-maneuver has fulfilled its purpose, he brings the Queen back
into the game. (To where it can play real role in the further play to come.)
17. Qf3!,
"The Queen returns to a stronger post, commanding the long diagonal." - Irving Chernev.

17...d6; (Center support.)
According to GM S. Flohr, ... (and GM A. Soltis!); this is not the best move here.

But as it looks VERY logical - AND the computer also chooses this move, - I cannot
possibly condemn it. (Also - Chernev did not criticize it either.)

The computer evaluation of this position - (after 17...d6;) after letting the dang thing run
ALL NIGHT LONG!!! ... (I have to sleep once in a while) ... is that Black is better. ("- 0.03")
This means Black is better, but only by three one-thousandths of a point. While this is hardly
enough to win a game, and probably over-emphasizes the problem of White's doubled
pawns, (and does NOT put enough weight on the long term value of the Bishop!); the final
conclusion here has to be that the position is currently equal. Also GM Nunn and FM Burgess
did NOT criticize this move at all! It is also interesting that no criticism of this move was made
in newspapers of this move either.

(Soltis awards a question mark here, - but for once - I'd have to say he was ... way off base!)

"17...c5; allows White to choose between playing on the Kingside, (18. g4); or the Queenside.
(18. Rb5.)" - FM G. Burgess.

<< A logical move - Black mainly prepares ...e5. (This is also the move most computer programs
choose here.) Soltis gives this move a question mark, but it is hard (impossible?) to agree
with his judgment here. >>

[ According to GM A. Soltis, it was better for Black to play ...c5!?; and ...Rc8.
    ("Better was ...c5 and ...Rc8." - GM A. Soltis.) But this does NOT stand up to analysis!
    For example: 17...c5; 18.Rb5!? Rc8?; This is bad here. (Black should play: 18...d6!; "="
     with a completely equal game, in this position.)   19.dxc5 Rxc5; 20.Qa8+ Rf8; 21.Qxa7, "+/"
    {Diagram?} with the better game for White.

   Even if Soltis would have chosen 18...d6; after Rb5, this is NOT clearly better than what was
   chosen in the game! (Not to be mean, but Soltis was probably not any stronger than Chernev.
   He certainly was NEVER in the world's Top Ten, as GM S. Tartakower certainly was.) ].

18. Re1, (Maybe - '!')
"It's work being finished on the Knight-file, the Rook moves to the center, to support a break
by P-K4." - Irving Chernev. (P-K4 = e4.)

[ One of my students suggested: 18.Kh1!?, with the idea of an eventual g2-g4. But the
outcome of such a plan is NOT completely clear. ].

18...Qd7!?; (Interesting, maybe even - '!')
Black removes himself (his Q) from the same file as the enemy Rook - common sense in chess.
(Maybe this move should be given an exclam!! And another point to this move is Black is
eyeing the White QRP now.)

"Black temporizes, as 18...P-K4; (18...e5;) 19. P-K4, (19. e4!) opens the lines to
White's benefit." - Irving Chernev.

A friend sent me a copy of some newspaper articles, (that had been copied onto micro-film) from
the time this game was played. The participants in this game apparently analyzed this game in detail,
and their thoughts were printed in a chess newspaper column of that time. And the consensus of all
the players was that 18...P-K4; may have been a little too risky.

I say all this only because Burgess (and Nunn) saw fit to give d6 a dubious appellation.

(Returning to our investigation of the main analysis line.)
19...dxe5; Black has to take. 20. Bxf5 Nxc4; "=" (Maybe - "+/="), with a fairly
equal game, according to Fritz 5.32. (This line was first generated in approximately in 1998
or 1999, on a friend's computer. So beware, because I think White is just slightly better here.)

I have since gone back and analyzed this line several times. {Once with Deep Junior!} White may
actually be a lot better here, or at least this is what the computer-driven analysis seems to indicate.

All this work does nothing ... or proves nothing ... except perhaps validate Black's decision in this line.
(Black decided, after a fair amount of time on his clock, not to play the move, 18...e5. Instead he
chose to remove his Queen from the e-file.) ].

I think we can safely dismiss those who would criticize Black's 17th or 18th moves!!

19. e4!,
White waits no longer - he blows open the center.
(This greatly increases the mobility of all of White's pieces, especially his Queen and Rook.).

19...fxe4;
Now White sets up a dual battery. (The e-file, and the b1-h7 diagonal.).

20...g6;
White has a nice position, but how does he improve his overall game?

With his next move, White consolidates/improves his Kingside.
21. g3!,
White fortifies the f-pawn, being in an open file. He also gains a square for his King.

FM G. Burgess notes that this is a preparation for a K-side attack, but there is very little
that Black can do about it.

[ Not as accurate is : 21.Rab1!? Nc6;   (21...Qxa4!?)   22.Rb5 a6; 23.Rg5, (This line was
    suggested by one of my students, and is very interesting.) Note that White has left his QRP
    unguarded. (23.Rb2!?)   23...Ref8; 24.g3 Nd8; "~" ].

21...Kf8; 22. Kg2, (Maybe - '!')
White prepares h2-h4, attacking the King-side.

"Cautious measures such as this, may have won battles for the Roman General, Quintus Fabius;
but against a player the stature of Capablanca, this amounts to simply waiting for the final blow
to descend." - Irving Chernev.

[ Black could have also played: 22...Qc6!?; 23.Qxc6 Nxc6; 24.c5 Re7; - I. Chernev.
(White probably is just a little better here.);

23. h4!, (Maybe - '!!') {Excellent!}
White has a majority on the right side of the board. He uses this majority to break down
his opponent's defenses and break up Black's Kingside for the endgame.

White also needs to act fast before Black has time to bring the Knight on a5 - which is
obviously very much 'out of play' - back into the game and consolidates his position.

[ 23.Qe2!? c6; 24.Rad1 Nb7; "=" White could have also tried: 23.Bf1, "="
or 23.c5!?; "~" ].

23...d5!?; (Hmmm.) {Diagram?}
This almost looks to be a panic reaction. A valid one though, considering what White could do
on the Kingside, given enough time.

"This leads to an exchange of the Queens, leading to a Rook-ending favouring White."
- Irving Chernev.

"Tartakower sees no sensible way to defend his Kingside, other than to exchange Queens.
However, this involves some concessions." - FM Graham Burgess. (& GM/Dr. John Nunn.)

<< Black decides the pressure on his position is great ... and perhaps in fear of a possible
K-side attack ... dumps the Queens, and opts for the end-game. >>

[ 23...Nxc4!?; 24.Bxc4 d5; 25.Bxd5 Qxd5; (25...exd5!?) 26.a5, "+/=" - Alekhine.

    23...Qc6!?; (Maybe - '!') - GM A. Soltis.
    But now 24.Qxc6 Nxc6; 25.c5!, "+/=" - LM A.J. Goldsby I
    (White threatens Bb5, winning.);

   Years of analysis has convinced me that Black's best move here may be: 23...c5!; "~" {Diagram?}
   as Black's only real try in this particular position. (White will generate a lot of pressure on the
   King-side. Black will have play on the other wing. A friend, with a new Pentium IV, allowed
   ChessMaster 8000 to look at this position all day while he was at work. The result was the
   computer felt Black was slightly better!  {About 15-25 one-thousandths of a point.}
   BUT ... a word of caution here. Even in the year 2002, the computers are VERY good at
   working out lines where there is an immediate {or so moves} tactical decision.
   They are MUCH less reliable at considering the extremely LONG-TERM possibilities of
   such a position! Especially in a long and complex ending!!!)

   I must confess that even with 23...c5; Black may still be lost - - - White will still have a very
   virulent K-side attack. BUT, if Black is lost, his overall SCHEME in the opening was refuted,
   and it is almost completely impossible to place the blame on any one move!! ].

White now decides to head for the endgame.
24. cxd5 exd5; 25. Qxe8+!,
The ending is the best bet for White.

25...Qxe8; 26. Rxe8+ Kxe8;
An ending has resulted where White has a good Bishop vs. a fairly limited Black Knight,
but 3 pawn islands vs. Black's two. Additionally, Black seems to have VERY promising
counterplay versus White's seemingly very weak (and backward) QBP.
(After the coming maneuver, ...R-f6 to c6.)

Black's pieces also have an iron grip on the c4-square. This is an excellent out-post square
for the Black Knight.

But Capa has seen much further than his opponent, and very accurately appraised this position.

<< With his next move, White opens lines on the K-side and isolates the Black
King's Knight-Pawn. >>
27. h5!, "+/="
White immediately pries open some lines on the King-side.

27...Rf6;
Fighting hard to stay in the game. (Black obviously cannot abandon the g6-square,
at least until his King moves up, because he would drop a pawn.)

[ 27...gxh5?; 28.Rh1 Kf8; 29.Rxh5 c6; (29...Kg8; 30.Rxd5, "+/-") 30.Rxh7,
wins a pawn for White. ].

28. hxg6 hxg6;
Black's 2 pieces are scattered on opposite sides of the board.

"How would you assess this ending? It may seem that c3-pawn is a very serious weakness, but it
turns out that the g6-pawn is just as easily attacked. Moreover, it is far easier for White to create
a passed pawn on the King-side than it is for Black on the queenside. Thus White should play
very actively, rather than trying to defend his queenside pawns." - FM G. Burgess.

White now seizes the best open file, then the 7th rank.
29. Rh1!, {Diagram?}
Occupying the open file, and headed for the seventh rank.

(Chernev and Soltis do NOT give this move an exclam. I give this move an exclamation point
because so many of my students - and in the early days ... when computers were much weaker,
most of the computer programs - choose the wrong move here.)

[ White could try: 29.Kf3!? Nearly ALL of my students pick the move: 29.Re1+!?, in this
position. 29...Kf7; "=" This is obviously MUCH less effective than what Capa chose!! ].

30. Rh7; (Really - '!')
White seizes the 7th row, here.
(Black's King will remain a prisoner for the rest of the game.)

Chernev here has a very long monologue here of the value of a Rook on the 7th Rank.
(Jan. 2002: It is interesting to note that even after 'thinking' for over 20 minutes on this
position, Junior 6.0 considers the game nearly level!!).

A.) "The Rook is in perfect position to attack to attack any pawns that have not yet moved -
those still standing on the second rank."

B.) "The Rook is prepared to attack any pawn that has already moved, by getting behind it
without loss of time. The pawn would then be under threat of capture, no matter how
many squares it advanced on the file."

C.) "The Rook's domination of the seventh rank confines the opposing King to the last rank,
preventing his participation in the fighting." (On the chess board.)

[ 30.Kf3!?, "=" 30.Rh8+? Kg7; 31.Rc8 Rc6; "-/+" Nimzo 7.32 rates this position as
nearly equal (!) and prefers the move: 30.Bb5!? "=" (Oct. 2001) ].

30...Rc6;
When I first saw this game, I was VERY concerned with how White would guard his QBP.
(Basically, he won't even bother.).

31. g4!, (Maybe - '!!')
A very fine move, and not at all obvious. White advances his K-side majority, and again
offers to trade his weakling of a c-pawn for Black's g-pawn.

<< White mobilizes his majority.
Look at this position. Does this honestly look like a totally won game for White?
(I don't think so.) >>

[ The computer program, Junior 6.0 examines this position for over 10 minutes and wants to
play the move: 31.Rd7!? "=" (Dec. 2001). 31...Nc4!; - GM A. Soltis.
Several of my students have suggested: 31.Bc2!? "=" ].

The Knight, which has been stuck on the edge of the board since move 16 (!), now ...
"hastens to get back into active play." (The quoted part from Chernev.)

[ Black should NOT play: 31...Rxc3?!; 32.Bxg6 Rc4; 33.f5!, (33.Kg3!? Rxd4; 34.Rxc7 Rxa4;
     35.g5 Ra3+; 36.Kg4 Nc4; 37.Kf5 b5; 38.Rc6 a5?!;   {Maybe Black should try: 38...Ke7!?;}
     39.Kf6! Re3; 40.Rc8+ Re8; 41.Rxe8#).   33...Rxd4; 34.Kg3 Rd3+!?; 35.Kf4, Rd4+;
    36.Kg5 Rxa4; 37.f6 Kg8; 38.Rxc7, "+/-" {Analysis Diagram.} (White is winning easily.)

     This variation clearly shows that by giving White two connected passed pawns, Black makes
     a severe and horrible mistake. The pawns are 'quicker,' i.e. they advance to the promotion
     square faster. Also, as the pawns advance, they will either gain time by checking the Black King,
     or creating mating threats - as in this last variation. ].

32. g5!?, (Probably - '!')
White - ...
A.) Avoids a premature exchange of pawns. White has also prevented Black from
playing ...g5. (White still is angling for connected passers.)
B.) Gains space. (White gets the g4-square for the White King.)
C.) Keeps Black guessing, keeps the pressure on, and maintains most of his options.
D.) White plans on Kf3-g4, slowly improving his position.
E.) Does NOT sweat Black's counter-play that starts with ...Nb2.

In addition to all of the above, Chernev notes that White has now FIXED the Black g-pawn
as a permanent target on the g6-square!

[ The computer program, Fritz 5.32 likes the move: 32.Kf2!?, "=" (Year, 2000);

32...Ne3+!?;
This is probably the best try for Black - indeed Black may even thought he was winning here.
(This is what was reported in several N.Y. newspaper columns the next day.)

[ If 32...Nb2!?; 33.Bc2 a5!; (33...Rxc3?; 34.Bxg6, "+/=")   34.Rh6!?, (34.Kf2!?, "=")
    34...Rxc3; 35.Bxg6, "~"   I prefer White in this endgame. As already has been seen, the
    long-range Bishop, active Rook, and connected-passed-pawns probably yield the better
    chances to White here.

[ Not 33.Kg1?! Nd1, "=/+" A totally different try is: 33.Kf2!?, This looks like a blunder,
     but it is not. I asked a friend, who has a very powerful computer and the brand-new
     program, Deep Junior; to test this game. He allowed the computer to analyze this game
     for over an hour, (In this position.); and this is the line the computer came up with.
     33...Nd1+; 34.Ke2 Nxc3+; 35.Kd2 Ne4+; 36.Bxe4 dxe4; 37.Ke3 Rc4; 38.Kxe4 a5;
     39.Kd5 Rxa4; 40.Rxc7 b5; 41.Ke5, "+/=" (Maybe - "+/") White maintains a slight
     advantage, because Black's King is still trapped by White's Rook. ].

33...Nf5;
If I were to rate this position, without knowing this game, I would have to say Black is better!!
(How does White guard his QBP?)

[ Chernev gives the (main) line: 33...Nd1; 34.Rh6, (34.Rd7!?) 34...Kg7; This seems
to be the best.

     ( Another try for Black is: 34...Kf7!?; 35.f5 Rxc3; 36.fxg6+ Kg8; 37.Ke2 Nb2;
       38.Bf5, "+/=" GM A. Alekhine - in the book of the tournament - stops here and says White
       is winning.   38...Nxa4; 39.Be6+, The most accurate.   (This is much better than the more
        natural looking: 39.Rh7 Rc6; 40.Ke3! Nb2; 41.Kf4 Nc4; 42.Re7 a5; 43.Be6+ Rxe6;
        (This is forced. 43...Kf8??; 44. g7+, winning.)   44.Rxe6 a4; 45.Kf5 Nd6+; 46.Ke5 a3;
        47.Kxd5 a2; 48.Re1 Kg7; 49.Ra1 Nb5; 50.Kc4 Nd6+; 51.Kd3 Kxg6; 52.Rxa2 Kxg5;
        53.Re2! Kf6 ; 54.d5 Kf7 ; 55.Re6 Nf5 ; 56.Ke4 Nd6+ ; 57.Kd4 Nf5+; 58.Ke5 Ng7;
        (58...Nd6; 59.Rh6! Nc4+; 60.Ke4! Nd6+; 61.Kf4! Ke7; 62.Rh7+ Kd8; 63.Ke5 Kc8;
        64.Ke6, "+/-"). 59.Rf6+ Ke7; 60.Rh6 Kf7; 61.Rh4, "+/-" and White should
        eventually win.)
       39...Kg7; 40.Bf7 Rg3; 41.Rh7+ Kf8; 42.Rh8+ Ke7;   (42...Kg7??; 43.Rg8#).
       43.Re8+ Kd7; 44.g7 Nc3+; 45.Kd2, "+/-" , {Diagram?} White should win easily.
       His g-pawn will cost Black a Rook. )

     35.f5 Nxc3;   (35...Rxc3?; 36.f6+ Kg8; 37.Ke2!, "+/")   36.Kf4! Ne4; 37.Bxe4,
      (37.fxg6!?)   37...dxe4; 38.f6+!, (Maybe - '!!')   (38.fxg6, "=")    38...Rxf6+;
      (This is forced. If 38...Kf7?; then 39. Kxe4, "+/")   39.gxf6+ Kxh6; 40.Kxe4 Kh7;
     41.Kd5! Kg8; 42.Kc6, (Maybe '!') The best?   (42.Ke6!? Kf8[]; 43.d5 g5; 44.Kd7 Kf7!?;
      45.Kxc7 g4; 46.d6 g3; 47.d7 g2; 48.d8Q g1Q; 49.Qe7+, ("+/-") with a won game for
      White. This is also probably just a transposition.).   42...g5;   (42...Kf7!?; 43.Kxc7 Ke6;
      44.d5+! Kxd5; 45.f7, "+/-")   43.Kxc7, ("+/-") Tartakower showed this entire line
     immediately after the game. (43.d5 g4 ; 44.Kxc7 , will just transpose.)   43...g4; 44.d5 g3;
     45.d6 g2; 46.d7 g1Q; 47.d8Q+ Kh7; 48.Qe7+ Kh6; 49.Qg7+! Qxg7+; 50.fxg7 Kxg7;
     51.Kb7 Kf7;   (51...a6; 52.Kxa6! Kf7; 53.Kxb6 Ke7; 54.Kb7, "+/-")   52.Kxa7 Ke7;
     53.Kxb6 Kd7; 54.Kb7, "+/-" and White wins.
     (A fantastic piece of analysis that the computers, in the year 2000, simply could not follow.).

34. Bxf5!?, (Probably - '!')
White must rid himself of this pesky Knight to make progress.

"Capablanca sees a rook ending as the simplest way to win. His Rook is very active, he has
a passed pawn, and he has seen a superb way to introduce his King into the thick of battle."
- FM Graham Burgess.

I must note that many Masters were watching this game as it was being played. Opinion was
quite divided as to the outcome of the game at this point. (Several Masters - such as Marshall
- felt that Black might even be better!)

[ The continuation: 34.Kg4 Ne3+; 35.Kf3 Nf5; 36.Kg4 Ne3+; is a draw by
repetition of position. ].

34...gxf5,
"Glancing at this position superficially, we see that White is about to lose a pawn. A deeper
look, however, ..." - FM Graham Burgess.
(He goes on to make several interesting comments about the course of the game.)

[ Black should NOT play: 34...Rxc3+?; 35.Kg4 gxf5+; 36.Kxf5 Rc4; 37.Ke5 Rxa4;
38.Rxc7 Ra2; 39.f5 Re2+; 40.Kf6, "+/-" and White is winning. ].

White now gives up two pawns with check in order to activate his King. Indeed, in order for this
to have worked; Capablanca must have planned this more than 10 moves in advance!

( OR "In a simplified ending, where Pawns are worth their weight in gold, he gives away
TWO Pawns! Moreover, he lets Black capture them WITH CHECK!!" - Irving Chernev.
{My emphasis.} )
35. Kg3!!, (Maybe - '!!!/!!!!') (Brilliant!)
"Le roi s'amuse! The King is headed for B6 (f6), a square from which he can assist the
White Rook in threats of mate, and also help the passed pawn take those last three steps."
- Irving Chernev.

'!' - Irving Chernev. '!' - GM A. Alekhine. '!' - GM A. Soltis. '!' - FM Graham Burgess.
(& GM/Dr. John Nunn.)

[ Most everybody I have shown this position to over the years, have chosen the move:
35.Rd7, "+/=" Additionally, many computers ... - over the years - have also chosen
this move!! ].

I should also point out that I analyzed this game on a friend's computer with the computer
program Fritz 4. (This was like in 1996.) The computer thought for over an hour, (in this position);
and it still considered this position to be ... completely winning for BLACK!!! (ha ha ha)

   [ 35...a6?!; 36.Kh4 Rxc3; 37.g6 c5; 38.dxc5 bxc5!?; 39.Kg5! Re3; 40.Kf6! Kg8;
     41.Rg7+! Kh8; 42.Rc7! Re8; 43.Kxf5! c4;   (Bad is: 43...Kg8?!; when Black is setting
      himself up for a fork as White advances his f-pawn.)   44.Kg5 Rd8; 45.f5 d4; 46.f6 Kg8;
      (46...d3; 47.Kh6 d2; 48.g7+ Kg8; 49.f7#)    47.Kh6, ("+/-") and White is winning. ].

[ Chernev gives the line: 36...Rc1; 37.Kh5! Rh1+; 38.Kg6 Rxh7; 39.Kxh7 c5;
40.g6, ("+/-") ... "and the Pawn crashes through." ].

37. g6!, (K invasion.)
"Once more Capablanca allows a Pawn to be captured with check! As compensation though,
Capa's King will gain an important tempo or two." - Irving Chernev.

37...Rxf4+; 38. Kg5, (Nice.)
"Gains time by attacking the Rook." - I. Chernev.

38...Re4;
Now White shows total disdain for material by not taking the f-pawn. As long as Black's Rook
can check from behind, White will leave the pawn there in order to shield his King.

[ "Capturing one more Pawn would be fatal: 38...Rxd4?; 39.Kf6 Kg8;   (On 39...Ke8;
      40.Rh8+ Kd7; 41.g7, and Black must give up his Rook for the Pawn.).   40.Rd7!,
     and mate follows." - Irving Chernev. ].

"Far superior to the petty 39. KxP. The disdained pawn will in fact be useful to White in acting
as a buffer against annoying checks."

"Capablanca's King is now in a dominating position, and faces Black with the threat of instant mate,
as well as 40. P-N7ch, followed by the promotion of the Knight Pawn." - Irving Chernev.

"Again highly instructive!" - FM Graham Burgess.
( He goes on to say pretty much the same things as Chernev has already noted. {Above.})
"It does not matter that Black has a mobile passed pawn, as White's threats are so immediate."
- FM G. Burgess.

'!' - Irving Chernev. '!' - GM A. Soltis. '!' - FM Graham Burgess. (& GM/Dr. John Nunn.)

The Black Pawn at f6 actually proves useful to White as a shield from unwelcome checks
by the Black Rook! >>

(I think prior to this game, this idea was very rare ... and may have been seen for the first time
here, in this game.) {A.J.G.}

<< The following series of moves are extremely precise by White. >>
39...Kg8; (Forced.)
White's active K+R combination totally tie Black up. White threatens mates, pawn promotions,
pawn grabbing, and just general mayhem. While this is a well-known device today, ...
it was ground-breaking, even revolutionary, at the time. !!! {A.J.G.}

[ If 39...Ke8?; 40.Rh8+ Kd7; 41.g7, "+/-" and White's pawn promotes.
Not 39...Rxd4??; 40.Rh8# ].

Now MOST players would just grab a pawn here.
40. Rg7+!, (Nice.)
Driving the Black King to a more exposed square.
(One where the g-pawn can advance with check?)

( Chernev does NOT give this move an exclam. But even Masters miss this finesse here.
Many of my students do NOT find this move here. So I think it definitely deserves an
exclam. {A.J.G.} )

41. Rxc7, (Almost - '!')
"Removes one pawn, and the renewed threat of mate will enable White to gather
in another." - Irving Chernev.

41...Re8; {Diagram?}
Now the Black Rook was forced back to the first rank, because of the mating threats.

Now that Black's Rook is stuck on his first rank, White captures the dangerous f-pawn.

42. Kxf5!,
Taking advantage of the fact that Black's Rook is now nailed to the first row.

Allowing the Black f-pawn to stay on the board increases Black's opportunities for counter-play.
(Chernev does NOT consider this move worthy of an exclam.)

[ 42.Rxa7?! f4; 43.Rb7?, (43.Rh7+ Kg8; 44.Rh1) 43...f3; 44.Rxb6 f2;
45.Rb1 f1Q+; 46.Rxf1 Rf8+; "-/+" ].

[ 42...Ra8; 43.Ke5, "+/-" Or 42...Rf8+; 43.Ke5 Re8+; 44.Kxd5 a6; 45.Rb7, "+/-" ].

43. Kf6!;
White threatens a mate and forces the Black Rook to return to the first rank.

(Again Chernev does NOT consider this move worthy of an exclam. But since so many of
my students play the wrong move here, I think it does deserve one. Timing is everything in chess.)

[ After the line, 43.Rxa7?! Rxd4; Black has much more counterplay than in the actual game. ].

43...Rf4+; 44. Ke5; (Almost - '!')
Noting the extreme activity of the White King, and the complete lack of any play by its
counter-part, Chernev notes that White is playing with an extra piece here.
("A King ahead," says Chernev.)

(A good rule of thumb is that an extremely active King, in an endgame where your opponent
has no activity for his monarch - is worth 2-4 points.) {A.J.G.}

(White's next move is also extremely accurate.)
45. g7+ Kg8;
This is forced.

[ 45...Rxg7?; 46.Rxg7 Kxg7; 47.Kxd5, "+/-" White has an easily won King-and-Pawn ending.

Chernev continues this line with: 47...Kf7; 48.Kd6 Ke8; 49.Kc7 Ke7; 50.d5, ("+/-") ...
"and the Pawn has clear sailing ahead." - I. Chernev. ].

46. Rxa7!,
Taking this pawn FIRST is the most accurate. (The d-pawn will not run away!)

[ 46.Kxd5 a5; 47.Rd7 Rg3; 48.Kc6 Ra3; 49.Kxb6 Rxa4; 50.d5, ("+/-") is probably still
winning for White, but is not as precise. ].

46...Rg1; 47. Kxd5 Rc1;
Black is trying to cut off the White King from running over and gobbling up his last pawn.

48. Kd6, (Probably - '!')
The King advances one square, so his Pawn can advance. Black doesn't even have a decent check.
(This type of improvement to ones' position almost never occurs to an amateur.)

50. Rc7 Ra1; 51. Kc6!,
Best. (White does not concern himself with trivialities.)

"This endgame provides a superb example of a number of important endgame themes: passed
pawns, rook activity, king activity, and an admirable avoidance of materialism when the initiative
is at stake." - FM G. Burgess.

53. Kb7 b5; 54. d7 Rd4; 55. Rc8+ Kxg7; 56. d8Q Rxd8;
57. Rxd8 b4; 58. Rd6,
So that the Rook can get behind the pawn.

58...b3; 59. Rb6 b2; 60. Rxb2, {Diagram?} ("+/-")
and as long as White knows how to mate with the King and Rook, he can win the game easily.).

<< Reti says, "No one has ever played these endgames with such elegant ease as Capablanca;
no one else has looked upon these technical difficulties so casually as a matter of course." >>
- Irving Chernev. (Who knew when to speak, when to quote someone else, and ... when to be silent!)

GM Salo Flohr called this game one of the great masterpieces of the chess board.

GM V. Smyslov said it was one of, ... "the finest Rook-and-Pawn end-games ever played."
(A {former} World Champion, who played more than a few fine endings of his own!)

GM A. Alekhine referred to this game as, ... "a monument to White's superlative skill in the ending."

<< One of the most convincing demonstrations of the power of a Rook in the seventh rank that was ever actually played
on a chess board. (As compared to possibly a composed problem.) >>

Irving Chernev, in his regular newspaper column, said this game was ... "in the Top 50 of the best chess games of all time."
(I would have like to seen that list!!)

This is game # 56 in Soltis's book, "The 100 Best." (By GM A. Soltis.) In addition to this, he showers this game with like 13 exclams!!
(And ... He rates this game just a few points from perfect in like 3 of 5 different categories!!)

This game is also in the book, [The Mammoth Book Of] "The World's Greatest Chess Games." (by GM's John Nunn, J. Emms, and of course, FM Graham Burgess.) (Game # 21.) In contrast to Soltis, FM G. Burgess, (& Nunn, who is supposed to have edited this book); only awards this game like 5 exclams, but four of these come in the ending. Since this book is about trying to pick the 100 best chess games of all time, this means these authors must also consider this game in the best one-hundred ever played.

Irving Chernev, in his book: "The Golden Dozen," calls this one of Capa's best games. I dare not disagree with perhaps the greatest writer who ever lived!

I have awarded like 20 (plus) exclams and one double-exclam in this game. Additionally, I have totally refuted the poor analysis of the many other annotators as concerns this game. Strategically, this is also a FLAWLESS game by Capablanca. (And a very good one by Tartakower, as well!)

After years of reflection and having gone over this game countless times, I think this game is one of the greatest games of all time. It is also one of Capa's very best efforts - that in itself is a HUGE statement. And this ending must easily be in the list of, "THE TEN BEST ROOK-AND-PAWN ENDGAMES EVER PLAYED!!!!!" - LIFE Master A.J. Goldsby I .

1 - 0

(Last {major} update: February 26th, 2003. Last edit on: Friday, August 02, 2013 02:45 AM . )

This game continues to fascinate chess players! See this month's chess magazine, (CL - October, 2004); for more details.

See GM Pal Benko's column, <Endgame Lab> ("Traces of Capablanca, Part 2"), page # 46. (Added 10/18/2004.)

This game is pretty much the full version of the game as it exists in my database.
(I have not shortened it for publication.)
If you would like a copy of that game to study, please contact me.

Since I posted this game ... even I have been surprised at the response it received. (It is definitely taken on a life of it's own and been really something of a phenomenon. I also used this game - or excerpts from it - in lectures on about a dozen different servers.) The response was immediate and overwhelming. This game has now been translated into the following languages, (with my permission): Spanish, French, German, Italian, Russian, Japanese and Chinese. (!)

Click HERE to return (or go to) to the "Home Page" of my biggest and best known site.

Click HERE to go to, or return to, my:
(formerly GeoCities) "Best (All-Time) [Chess] Games Page."

Click HERE to go to my "Best Short Games Page."
(If you liked this game, you will enjoy the many games that are available on this page.)

Click HERE to go to, or return to, my: "Best All-Time [Chess] Moves Page."

J. Capablanca - S. Tartakower; NY, 1924

GM Jose R. Capablanca (2740) - GM Savielly G. Tartakower (2650) [A80] The New York International Chess Tournament New York, NY (Rd. # 6), 1924

[A.J. Goldsby I]

1 - 0

J. Capablanca - S. Tartakower;
NY, 1924

GM Jose R. Capablanca (2740) - GM Savielly G. Tartakower (2650)
[A80]
The New York International Chess Tournament
New York, NY (Rd. # 6), 1924