More On Exchange Sacrifices

     Here is some more on the sacrifice of the exchange based on instructions from that master of attack, Rudolf Spielmann. 

     Spielmann observed that the absolute values are what form the basis on which most exchanges are made, i.e. Q= 9, R=5, B and N = 3 and P=1. However, it is the relative value that is the decisive factor for positional play and especially for sacrifices. To wit: 

 

1) The simpler the position, the more the absolute value carries weight. 

2) The more complicated the position, the more the relative value of the pieces gain in importance. 

 

     When it comes to the sacrifice of the exchange it is never an exactly even transaction so you either win the exchange or you lose it. Exchange sacrifices are dependent on specific positional and tactical factors. 

 

Spielmann offers the following helpful definitions: 

 

1) Any voluntary loss in material counts as a sacrifice. 

2) The sacrifice of the exchange is when a Rook is given up for a minor piece and a Pawn. 

3) The term sacrificing the exchange when applied to situations where a Rook is given up for a minor piece and two Pawns is incorrect; it should be wins two Pawns for the exchange. 

4) The sacrifice of the exchange can serve any purpose. It can be a sacrifice for: 

     a) Development 

     b) An obstructive sacrifice, a sacrifice intended to blockade a square, file, rank or diagonal. 

     c) A sacrifice designed to expose the opponent's King. 

 

     In all cases, the sacrifice of the exchange is designed to improve the position of the minor pieces. 

     Here is a case in point taken from the game Spielmann vs. Tarrasch that was played in Carlsbad 1923. Spielmann gave this game in his classic The Art of Sacrifice in Chess. A great attacker and sacrificial player, in this book he classified various tactical motifs, but as with almost all of those old books from pre-engine days, analytical errors abound in the examples. Still, there is a lot to be learned from the book and even if his evaluations and play in the following game were not always correct, they worked against the great Tarrasch which says something...we can learn ideas from them and maybe use theme in our own games.

Rudolf Spielmann - Siegbert Tarrasch

Result: 1-0

Site: Carlsbad

Date: 1923

King's Gambit Declined

[...] 1.e4 e5 2.f4 ♗c5 Statistically, at least in my database, this move produces much poorer results for black than accepting the game. White has a lot of ways to counter this move and many times black will end up with a much worse position than if he had accepted it. If black wants to decline the gambit then the Falkbeer (2...d5) is a better choice. The idea is that the B prevents white from castling. 3.♘f3

3.fxe5 Is a trap that loses for white. 3...♕h4+ 4.g3 (4.♔e2 ♕xe4#) 4...♕xe4+ wins the R.

3...d6 4.c3 Other options are 4.Nc3 and 4.Bc4, but the text is considered slightly better. 4...♗g4 This pin is somewhat annoying and the method white uses here to counteract it originated with Frank Marshall. 5.fxe5 dxe5 6.♕a4+ ♗d7 7.♕c2 ♘c6 8.b4 This is the point as black's next move is forced. 8...♗d6 Necessary, else 9 P-N5 wins the King Pawn.

8...♗b6 9.b5 ♘a5

9...♘ce7 is much worse because after 10.♘xe5 ♘f6 11.a4 white's position is even better than after 9...Na5

10.♘xe5 winning the e-Pawn.

9.♗c4

9.b5 isn't so effective now as seen in Huschenbeth,N (2532)-Azarov,S (2635)/Greensboro USA 2014 9...♘a5 10.d4 and now black could have equalized with 10...c6. Instead he played the much weaker 10...♕e7 and white got the upper hand after 11.♗d3 f5 12.♗g5 ♘f6 13.♘bd2 h6 14.♗xf6 gxf6 15.O-O

9.♘a3 does not give white much. 9...f5 10.d3 ♘f6 11.♘c4 O-O 12.a4 a6 13.♗e2 h6 14.O-O is equal. Bakhtiari Kish,M (2056)-Dilanyan,G (2296)/Erevan 2016(21)

9...♘f6 10.d3 ♘e7

10...h6 11.O-O O-O 12.a4 a6 13.♘bd2 b5 14.♗b3 equals. Popovych,O (2305)-De Lange, D (2210)/Gausdal 1982

10...♕e7 11.O-O O-O-O 12.a4 a5 13.b5 gave white a huge advantage in Bronstein,D-Panov,V/Moscow 1947

11.O-O ♘g6 12.♗e3 This permits black to obtain counterplay. 12.a4 was correct. 12...b5 13.♗b3 a5 14.a3 axb4 15.cxb4 Black's Q-side advance has succeeded in forcing white to play this which has resulted in black achieving full equality. 15...O-O

15...♗xb4 is refuted by 16.♗xf7+ ♔xf7 17.♕b3+ ♔e8 18.♕xb4 Although the material is equal white is much better.

16.♘c3 c6 17.h3 ♕e7 18.♘e2 ♗b8 The purpose of this move is to exchange Bs after ...Ba7 in the hope of establishing a N on f5, but the maneuver is somewhat labored. The immediate 18...Nh5 was simpler. 19.♔h2 ♗a7 20.♗g5 h6 21.♗xf6 ♕xf6 22.♘fd4 ♕d6 23.♘f5 ♗xf5 24.♖xf5 ♘f4 25.♖f1 Up to this point black has defended himself ably and picked up a positional advantage in the better P-formation because white's Ps on a3 and d3 are backward and may become weak. For his part white has built up strong pressure on the f-file which Tarrasch seems to underestimate. (Spielmann) Actually, white's last move gives black a slight advantage, but he must play accurately because the slightest slip will give white tactical chances. It doesn't matter how great a positional advantage you have, miss a tactical shot and the positional advantage means nothing!

25.♘xf4 was correct. Then after 25...exf4 26.e5 ♕d4 27.♖f1 ♖ad8 28.♖5xf4 ♕xe5 29.g3 the chances would be about even as black must guard against the attack on f7.

25...g6 Her Spielmann makes the incorrect claim that it would be better to decline white's offer of the exchange implied by his last move and that black had little choice but to play 25...Ne6. The text move "wins" the exchange at the cost of enabling white's attack, hitherto directed only against f7, to spread over the whole K-side. (Spielmann)

25...♘e6 This move does result in an equal position so Spielmann is correct in that claim. 26.♕a2 ♖ae8 is completely equal.

26.♖1xf4 exf4 27.e5 Black can claim a slight advantage here...provided he finds the best moves...it is this difficult task that makes the games of great attackers like Spielmann, Nezhmetdinov, Tal and others so entertaining...in the complications their opponents often went astray. 27...♕e7

27...♕c7 With this move black would have kept his small advantage in hand. 28.♖f6 ♔h8 (28...♕xe5 A poisoned P! 29.♖xg6+ ♔h7 30.d4 wins) 29.♕c3 ♗e3 30.♖xc6 ♕d8 and black is doing OK.

28.♖f6 Spielmann has completely misjudged the position, not only did he do so during the game, he did so when he wrote the book! He wrote - "The powerful establishment of the R on this square is the point of the sacrifice. There are now many threats, above all 29.d4 and only then capture at g6 by either the R or Q. True, Black remains with two Rs for the Q, but his K-side is so critically weakened that successful defense is, in the long run, impossible. Yet this would be better than the defense which Bback actually selects." That's all wrong as any engine will tell you. Stockfish puts black's advantage at nearly 4 Ps, Komodo about one P while the ancient Fritz 5.32 puts it at about 3/4 of a P.

28.♕xc6 This move, unmentioned by Spielmann, would have drawn which is white's best course. 28...gxf5 29.♕g6+ ♔h8 30.♕xh6+ ♔g8 31.♕g6+ ♔h8 32.♕h6+ ♔g8 33.♕g6+ ♔h8 34.♕h6+

28...♔g7 I won't even bother quoting Spielmann here because his note is so far from reality! However, this is the losing move after which Stockfish evaluates the position as being almost 6 Ps in white's favor which is, obviously, a clear win.

28...♔h8 is another story...black is winning! 29.♕c3

29.♕xc6 ♕xe5 30.♖xf4 ♖ae8 This is even better than taking the N. Black has a won game in either case.

29...♗e3 The sacrifice on g6 is no longer a threat. 30.♖xg6 fxg6 31.e6+ ♔g8 32.♕xc6 ♖xa3 and the game is practically over.

29.d4 And now it's true...white is winning. 29...♗xd4 Tricky! Of course, white cannot capture the B at once. 30.♗xf7

30.♘xd4 White must leave the B alone! 30...♕xe5 There is nothing white can do. His best line is 31.♖xf7+ ♖xf7 32.♗xf7 f3+ 33.♔h1 fxg2+ 34.♔xg2 ♔xf7 35.♕xc6 ♖d8 36.♘f3 He needs the N to defend his K. 36...♕e2+ 37.♔g3 ♖d3 and all white has left is some meaningless checks.

30...♗xe5 This allows a mate in two, but he was lost anyway.

30...♖xf7 31.♕xg6+ ♔f8 32.♘xd4 ♖xf6 33.exf6 ♕f7 34.♕xh6+ ♔e8 35.♘f3 ♕f8

35...♖xa3 36.♘g5 ♕f8 37.♕g6+ ♔d7 38.♕f5+ and wins...the K has no safe retreat.

36.♕g6+ ♕f7 37.♕e4+ ♔d8 38.♕xc6 White has a won ending. For example... 38...♖xa3 39.♘e5 ♕h7 40.♕d6+ ♔c8 41.f7 forces mate in 11 moves.

31.♕xg6+ Facing mate next move, Tarrasch resigned. An entertaining game!

Powered by Aquarium

The Exchange Sacrifice in the Sicilian

     This sacrifice of the exchange is not a great sacrifice because the exchange is, materially speaking, worth less than a piece. That is, a R is worth 5 Ps while a piece is worth 3 Ps, so the difference in point value is two Ps. And, there are many positions in which the Rook is unable to exercise its full power, or where a well-placed N or B may be exceptionally powerfully posted. 

     The sacrifice of the exchange is often a way of trying to exploit special features in a position; in those cases the superficial material valuation of five points against three is meaningless. Of course the sacrifice of the exchange is even better for the side giving up the Rook if that side also gets one or more Pawns. 

     When deciding on whether or not to sacrifice the exchange masters rely mostly on calculation and the feel for the position rather than general considerations. However, here are a couple of rules of thumb that can be used as a guideline by the rest of us.

A) You can get an attack or a passed Pawn 

B) Sacrifices that weaken the opponent’s P-structure or which lead to a strong initiative will likely give winning chances. 

 

      When sacrificing the exchange it is wise to take the following generalizations into consideration:

 

 A) With only a Pawn for the exchange the chances are it's an even exchange. 

B) With just a R for N (no Ps) your opponent may be able to return the exchange and win a P himself. 

C) If your opponent does not have a B, two Bs can sometimes equal a R and N. 

D) If the Ps are all on one side of the board, there is less danger of losing the ending for the side that is down the exchange. 

 

      In the Sicilian Defense a black R often finds itself on c8 from where it puts pressure along the half-open file and so white, especially if he has castled Q-side, must always be on the lookout for a positional sacrifice of the exchange by black. 

     The ...Rxc3 exchange sacrifice is one of the most well known tactical patterns. This sacrifice can be used to serve several different purposes. Actually, although it is most dangerous if white has castled Q-side, ...Rxc3 can be played regardless of which side white has castled on as it can be used for either immediate or long-term gains. 

      In positions where the players have castled on opposite sides it is likely to lead to very sharp positions. In this case both sides must balance their attack on the enemy K with the defense of their own. To do this requires experience, knowledge and intuition which is something we non-masters do not have to a great degree!

     For the serious student who wants to gain knowledge of the exchange sacrifice studying a book such as The Exchange Sacrifice, A Practical Guide by Sergey Kasparov, a GM from Belarus, might be worthwhile. The nice thing about the book is the over 200 illustrative games. 

     Efim Geller is best remembered for his tactical ability and original attacking style which characterized his early career. In later years he became a more rounded player and an expert on openings. 

     He knew all about the ...Rxc3 sacrifice of the exchange in the Sicilian and we see him carry it off against Isaac Boleslavsky in their amazingly complicated game in the Zurich Candidates Tournament in 1953. 

     In his classic book on the tournament David Bronstein gave the impression that Geller had the game in hand from almost the beginning. Of course, with the benefit of engines we see that was not the case. In fact, the complications were so great that in going over this game I ran more Shootouts to get some idea of the results than I have in any game I have recently posted! As a reminder, Shootouts allow you to use one or more engines to play out the rest of the game at different ply depths with the sides being changed after each game.

Isaac Boleslavsky - Efim Geller

Result: 0-1

Site: Candidates Tmt, Zurich

Date: 1953

Sicilian Dragon: Yugoslav Attack

[...] 1.e4 c5 2.♘f3 ♘c6 3.d4 cxd4 4.♘xd4 ♘f6 5.♘c3 d6 6.g3 ♗g4 An interesting move that forces Geller to play 7.f3. The opening ends up with Boleslavsky playing the Rauzer Attack against Geller's Dragon. 7.f3 ♗d7 8.♗e3 g6 9.♕d2 ♗g7 10.O-O-O O-O 11.g4 ♖c8 12.♔b1 ♘e5 13.h4 Let's let Bronstein explain this position! White has a well placed N on d4; from there it can meet black's counterattacking ...Qa5 with Nb3 which force the Q to retreat. Additionally, white has started his P-storm first, having already played g3-g4 and h2-h4. Note that the position of white's K has no weakness while black's P on g6 is a focal point for white's advancing Ps. For his part, although black will be starting his P advance later than white, black already has open against white's K. Also, his centralized N t is very well placed as it attacks the weak link in white's P-chain, the P on f3 which is the only support of white's whole shaky P-chain. Black controls, and may soon occupy, the important square c4. Meanwhile, white still does not control one square in the immediate area of black's K. Finally, the powerful B on g7 may give rise to dangerous tactics on the long diagonal. All of this adds up to a position that is in a state of dynamic equality and great skill is necessary in order to maneuver the entire mass of Ps and pieces while simultaneously countering the enemy's operations. 13...b5 This prepares both ... b4 and ...Nc4. After ...Nc4 white must avoid Bxc4 because black would recapture, not with the R, but with the P because th e semi-open b-file would be more dangerous for white than the semi-open c-file. 14.♗h6 Taking the b-pawn would be madness, of course. And h4-h5 is ineffective so long as Black's pieces solidly defend the squares h7 and h8; so first white wants to trade off one of the most important defensive pieces.

14.♘cxb5 is not advisable. 14...♘xf3 15.♘xf3 ♘xe4 16.♕g2 ♗xb2 17.♔xb2 ♗xb5 18.♗xb5 ♕a5 19.c4 (19.♗d3 ♖b8+ mates in 3.) 19...♕c3+ 20.♔b1 ♕xe3 This position is very complicated and five Shootouts were all drawn. However, it required very careful defense on white's part to hold the draw and in practice the slightest slip would be disastrous.

14...♗xh6

14...♗h8 trying to use the B on the long diagonal fails. Jst a sample of what could happen... 15.♗xf8 ♔xf8 16.h5 g5 17.♕xg5 ♖xc3 18.bxc3 ♘xf3 19.♘xf3 ♘xe4 20.♕f4 ♗c6 21.♘g5 ♘xg5 22.♕xg5 ♗xh1 23.♗d3 ♗f6 24.♕xb5 ♗f3 25.♖f1 ♗xg4 26.h6 White has a won ending and won 5 Shootouts. Here is the rest of the game at 17 plies. 26...♗d7 27.♕b7 ♗xc3 28.♗xh7 e6 29.♗g6 f5 30.♕b3 ♕f6 31.h7 ♕xg6 32.♕xc3 ♕xh7 33.♕f6+ ♕f7 34.♕d4 ♕g7 35.♕xd6+ ♕e7 36.♕e5 ♕g7 37.♕e3 a5 38.♖g1 ♕f6 39.♕a3+ ♔f7 40.♕h3 ♔e7 41.♕h7+ ♔d8 42.♖d1 ♕e7 43.♕h8+ ♔c7 44.♕e5+ ♔b6 45.♕d4+ ♔b5 46.♕xd7+ ♕xd7 47.♖xd7

14...b4 is met by 15.♘d5 ♘xd5 16.exd5 ♗xh6 17.♕xh6 and white's attack will come first.

14...♘c4 15.♗xc4 bxc4 16.♗xg7 ♔xg7 17.h5 Here, too, white's attack comes first.

14...♖xc3 is also unsatisfactory. 15.♗xg7 ♖xf3 16.♗xf8 ♕xf8 17.♘xf3 ♘xf3 18.♕e3 ♘e5 19.g5 ♘xe4 20.h5 (20.♕xe4 ♗c6) 20...♘c5 21.hxg6 ♘xg6 22.♗h3 ♗c6 White, being the double exchange ahead, is winning. Here is how Stockfish accomplished it. 23.♖h2 ♗e4 24.♗g2 ♗xg2 25.♖xg2 ♕c8 26.♖h1 ♕e6 27.♕f3 ♕e4 28.♕h3 ♘f8 29.♖f2 ♕e6 30.g6 fxg6 31.♕f3 ♘cd7 32.♖hf1 ♘f6 33.♖e2 ♕d7 34.♖fe1 ♔f7 35.♖xe7+ ♕xe7 36.♖xe7+ ♔xe7 37.♕b7+ ♘8d7 38.♕xb5 h5 39.♕g5 ♔f7 40.a4 ♘e5 41.♕e3 ♘c6 42.♕b3+ ♔e7 43.♕b7+ ♔d8 44.♕xc6 ♘e8 45.♕a8+ ♔d7 46.♕xa7+

15.♕xh6 ♖xc3 h4-h5 was a real threat now, so this exchange of rook for knight, which also breaks up the king's cover, was practically forced. 16.bxc3 ♕a5 The natural attacking continuation. 17.♕e3 White's king position is loosening, and his queen must now return to the defense. On the other hand, the position is still in balance, since White has the exchange as compensation. 17...♕a3 Another interesting position, especially these days with engines at our disposal. In annotating this game in his great book on the Zurich tournament, Bronstein makes no comment on the next couple of moves, but Stockfish is critical of them. In fact, many of Bronstein's remaining comments are erroneous, but let's not be too critical because he did not have Stockfish and Komodo. If anything, his comments and the engine's analysis show the complexity of the rest of the game!

17...♖c8 Has been tried in later games. 18.♔b2 Less precise is 18. Nb3 18...b4 19.cxb4 ♕xb4+ 20.♕b3 ♕c5 21.♗a6 and white soon won. Escandell Mari, J (2158)-Sauleda Roig,J (2172)/ Barcelona 2005

17...♘c4 This is totally unsatisfactory. 18.♗xc4 bxc4 19.h5 ♖b8+ 20.♔a1 e5 21.♘e2 ♖b6 22.♖b1 ♖a6 23.♖b2 and white has a winning position.

18.h5 Bronstein appears to have based his annotations to a large extent on the results. In auto-annotation mode this move gets two ? from the engines which swing the evaluation to black's favor after this move. In order to maintain a slight advantage white had to take a moment to make some defensive moves.

18.♗e2 ♖c8 19.♕c1 ♕xc3 and now 20.h5 with the better position, but only slightly!

18...b4 Logical, but the players and Bronstein missed a tactical shot that secures the advantage.

18...♘xf3 19.♕xf3 ♗xg4 20.♕e3 ♗xd1 Black has won back the exchange and is even a couple of Ps ahead. Besides that he has an adequate defense for his K.

18...♖c8 was a less good alternative for black. 19.♘b3 a5 20.♗e2 b4 21.cxb4 a4 22.b5 axb3 23.cxb3 ♗e6 24.♖c1 ♖a8 25.♖c2 In the long run white's position will prove superior.

19.♕c1 If black is to have any hopes of attacking he cannot exchange Qs. 19...♕xc3

19...♕xc1+ 20.♖xc1 bxc3 21.♔a1 ♖b8 22.♖b1 ♖xb1+ 23.♔xb1 and white should prevail.

20.♕b2 Even better was the march of the Ps starting with 20.h6

20.h6 a5 21.g5 ♘h5 (21...♘e8 22.♕d2 ♕c7 23.f4 and black's forces are beaten back.) 22.f4 ♘g4 23.♗h3 f5 24.exf5

20...♖c8 Bronstein comments: Black refrains from 20... Qe3, which could result in a repetition of moves. Black has no advantage and after the exchange of Qs his game becomes perhaps even a bit inferior. Bronstein's comment is correct, but what he failed to notice is that with 20... Nxf3 black fully equalizes.

20...♘xf3 21.♕xc3 (21.hxg6 ♘d2+ 22.♔c1 ♘fxe4 23.gxh7+ ♔h8 24.♗d3 ♘f2 is also equal.) 21...bxc3 22.♘xf3 ♗xg4 This same idea was seen in the note to black's 18th move.

21.hxg6 This more or less forces black to trade Qs and it should be nted that at this point it is white who stands slightly better. 21...♕xb2+ (21...fxg6 22.♕b3+ ♔g7 23.g5 is also unpleasant for black.) 22.♔xb2 hxg6 23.a3 Bronstein comments that this is an inaccuracy of such magnitude that it loses the game, but that is more fiction than fact. True, it is an inaccuracy, but all it does is let his position slip from better to equal.

23.g5 keeps white slightly better after 23...♘h5 24.♖h4 a5 25.f4

23...bxa3+ 24.♔xa3 Bronstein recommended 24.Ka2, but Boleslavsky played this based on a miscalculation. Even so, there is nothing wrong with the move played. 24...♘xf3 Boleslavsky saw this. 25.♘xf3 ♖c3+ When he played 24.Kxa3 Boleslavsky envisaged that he could meet this with 26. Rd3 guarding the N! Still, the position is equal. 26.♔b2 ♖xf3 27.e5 It is here that Boleslavsky starts to go astray.

27.g5 is the correct move. 27...♘xe4 28.♗g2 ♘xg5 29.♗xf3 ♘xf3 30.♖a1 ♗f5 31.♖xa7 with an incredibly unbalanced position. Five Shootout games were drawn.

27...♘xg4 With this move white's position cracks. 28.♗e2 ♖f2 29.♗xg4 ♗xg4 30.♖df1 But, this is the move that loses the game.

30.♖d4 would have allowed black to play on. 30...♗f5 31.exd6 ♖xc2+ 32.♔b3 exd6 33.♖xd6 Black is slightly better, but it's time for another Shootout! Black won one game and 4 were drawn.

30...♖xf1 31.♖xf1 dxe5 As Bronstein correctly pointed out, no R can fight a B and four Ps!. 32.c4 ♔f8 33.♖a1 Going after the a-Pawn costs time which black will use to advance his horde of Ps on the K-side.

33.c5 would not have saved the game though. 33...♔e8 34.c6 f5 35.c7 ♔d7 36.♖c1 ♔c8 37.♖c6 f4 38.♖xg6 ♗f5 39.♖g5 e6 and wins.

33...♗f3 34.c5 g5 35.♖xa7 g4 36.♖a3 ♔e8 37.♔c1 f5 38.♔d2 f4 39.♖a6 g3 40.♔e1 ♗e4 One assumes the time control has been reached and so white resigned. An extraordinarily complicated game!

Powered by Aquarium