The structure e-, f- and g-pawns against f- and h-pawns arises very often in practice, especially in 4 versus 3 rook endings, when the defender allows the exchange of h- und g-pawns (which is generally favourable for the attacking side). The game Duras-Capablanca 1913 (with col. rev., Capablanca had the 3 pawns) is the first practical example of this kind of endgame which showed the attacking possibilities of the stronger side. Together with the similar ending Capablanca-Yates 1930 (see the thread “Capablanca-Yates” in this forum) it is considered as a classic. We begin the analysis from the 60. move of Duras when the e-pawn has already advanced to the fifth rank (later on we will consider positions with the e-pawn on third and forth ranks). The endgame theory has correctly established that this pawn structure with the central pawn on e5 is winning for White.

D. 1



Duras-Capablanca 1913 (rev. col.)

All pieces of White are optimally placed. Once White manages to push the bishop pawn to f5 Black’s fate is sealed.

1...Rb4+

This is the only move to put up some resistance. After 1…Rb1 2.f5 the game is practically over. With the pawn on f5 and the rook on seventh or eighth rank Black does not have any counter-play.

2.Rd4 Rb1

D. 1.-1



Capablanca played here 3.f5 which has been regarded by the theory for a long time as a natural move. But 1956 N. Kopayev came to the conclusion that it is a result-changing error (Shakhmatny Biulleten, n. 1956, p. 266-267). All other later analysts who dealt with this ending, including M. Euwe in the last edition of his Het Eindspel 3 (together with van Wijgerden, Utrecht/Antwerpen 1981, p. 128), adopted the conclusions of Kopayev. In the thread “Capablanca-Yates” of this forum I have also uncritically accepted the analysis of Kopayev. After a thorough analysis of this ending I think now that 3.f5 does not throw away the win. The main reason of the negative evaluation of 3.f5 by Kopayev was that the immediate advance of the pawn gives the Black king the possibility to come out to the e-file and to take active part in the game. Although this assessment of Kopayev remained uncontested in the theory, in nearly all practical games White played 3.f5. While I will prove here that 3.f5 wins it should be stressed that it is not the optimal winning line. 1956 Kopayev showed the most convincing path to victory. He suggested pushing the f-pawn to the 5th rank after transferring the rook to the eighth rank, not earlier. In that way the Black’s king which is restricted in his mobility not only by White pieces but also by his own pawns is shut out of the game. See D. 2 for illustration of that plan. 

3.f5?!

D. 1.-2



Now White threatens 4.f6+ Kh7 5.Rd7 Kg6 6.Rd8 Rb4+ 7.Kf3 Rb3+ 8.Kf2 Rb2+ 9.Kg3 Rb3+ 10.Kh4 Rb1 11.Rg8+ Kh7 12.Rg7+ Kh8 13.Rxf7+-.
6…Rb7 also does not help. After 7.Rg8+ Kh7 8.Rg7+ Kh8 9.Kf5 Ra7 White sacrifices his rook which “removes the last obstacle in the way of the pawns” (K. Mueller). 10.Rxf7 Rxf7 11.e6 Kg8 12.exf7+ Kxf7 13.Ke5 Kf8 14.Ke6 Ke8 15.f7+ Kf8 16.Kf6 (1-0 Kulaots-Fedoseev 2012 with rev. col.) 16…h5 17.g5 (1–0 Van Beek-Span 2007 with rev. col.).

Duras played here 3…Rg1? (I.) which loses quickly. Kopayev’s move 3…Kf8 (II.) offers greater resistance but against accurate play it also does not save the Black. In practical games Black tried many other moves but neither of them is sufficient to reach a draw (III.-VI).

I. 3…Rg1? 

D. 1.-3



4.Kd5?

Capablanca fails to exploit his opponent’s mistake. Kopayev’s move 4.Rd7! wins here faster:
a) 4...Kf8 5.f6 Re1+ 6.Kd5 Rd1+ 7.Kc6 Rc1+ 8.Kb5 Rb1+ (8...Ke8 9.Re7+ Kf8 10.Ra7 Ke8 11.Ra8+ Kd7 12.Rf8 Ke6 13.Re8+ Kd5 14.e6) 9.Kc4 Rc1+ 10.Kb4 Rc8 11.Kb5 Re8 12.Kc6 Rxe5 13.Rd8+ Re8 14.Rxe8+ Kxe8+- 
b) 4...Rxg4+ 5.Kf3 Rg5 6.Kf4 Kf8 7.f6 Ke8 (7...Kg8 8.Rd8+ Kh7 9.Rf8 Kg6 10.Rg8+ Kh5 11.Rg7) 8.Re7+ Kf8 9.Ra7 Ke8 10.Ra8+ Kd7 11.Rf8 Ke6 12.Re8+ Kd5 13.Re7 Rg8 14.Rd7+ Kc6 15.Rxf7+-
In the game Sideif-Zade-Tavadian 1979 (rev. col.) Sideif-Zade played 6…Kg8 and resigned after 7.Rd8+ Kg7 8.f6+ Kh7 9.Rf8 Kg6 10.Rg8+ Kh7 11.Rxg5 hxg5+ 12.Kxg5 Kg8 13.Kh6 Kf8 14.Kh7.

4...Kf8!

Black seizes the opportunity to activate his king, heading for e7. But contrary to theory books,
Black can not save this position.   

5.Kd6 Ra1

D. 1.-4



Capablanca played here 6.e6? (2) which lets the win slip. Kopayev believed that 6.f6 (1) is also draws but in fact White wins.

1) 6.f6!

D. 1.-5



This move has been underestimated by the theory. Although Kopayev analysed it, he failed to asses correctly the position after 6.f6. White could have played this move also after 4…Kf8. 

6...Ke8

Kopayev analysed here 2.Rb4 (B) which is inferior to 2.Rc4 (A).

A) 7.Rc4

White has to keep his king on c6 or c7 in order to prevent the release of the Black king via d7. White places his rook on c-file and protects his king from the checks along this file.

7...Rd1+ 8.Kc6

D. 1.-6



This is the best place for the White king – he is protected from the checks from behind and he does not allow the Black king to come out to d7. White’s plan is to put the rook on c5 and, after having protected the e-pawn, to transfer the rook to a5. When the rook is on a5 (or b5) the checks from behind are not dangerous.
8.Kc7 is inferior as White wants to get the position of the main line after 3…Rg1 4.Kc7 with Black to play.

8...Rg1

Trying to prevent the move Rc5. 3...Ra1 offers little resistance: 4.Rc5 Kd8 5.Rd5+ Ke8 (5...Kc8 6.Kd6 Kd8 7.e6+-) 6.Rb5 Rc1+ 7.Kd5 Rd1+ 8.Ke4 Re1+ 9.Kd3 Kd8 10.Rb8+ Kd7 11.Rb7+ Ke8 12.Re7+ with transposition to the main line after 19.Re7+.

9.Kc7 Ra1

This is forced because staying on g-file shortens the distance between the rook and White king. White gains a tempo and transfers his rook to f8. 9...Rg2?! 10.Ra4 Rc2+ 11.Kd6 Rd2+ 12.Kc5 Rc2+ 13.Kd4 Rd2+ 14.Ke3 Rd5 15.Ra8+ Kd7 16.Ke4 Rd1 17.Rf8 Re1+ 18.Kd5 Rd1+ 19.Kc4 Rc1+ 20.Kd3 Ke6 21.Re8+ Kd5 (21...Kd7 22.Re7+ Kd8 23.Rxf7+-) 22.e6+-.

10.Rc5 Rg1 11.Ra5

Now that e-pawn is protected the rook can leave the c-file.

D. 1.-7



11…Rc1+ 12.Kd6 Rd1+ 13.Kc5 Rc1+ 14.Kd4 Rd1+

This is the most stubborn defence. After 14...Rc8 White’s task is easier - 14...Rc8 15.Ra1 Kd7 16.Ke4 (16.Rh1? Rg8! 17.Rh4 Ke6=) 16...Rc2 (16...Ke6 17.Ra6+ Kd7 18.Kf5+-) 17.Ra7+ Ke8 18.Ra8+ Kd7 19.Rf8 Re2+ 20.Kd4 Rd2+ 21.Ke3+-. 

15.Ke4

15.Kc3?!  Rc1+ 16.Kd2?! is bad because after 16…Rf1 Black threatens to attack the g-pawn with Rf4. White has to play 17.Ke3! and after 17...Kd7 18.Ra6! Re1+ 19.Kf4 Rf1+ 20.Kg3 Rg1+ 21.Kf3 Rh1 22.Rd6+ Ke8 23.Rd3 Rf1+ 24.Ke4 Rg1 25.Kf4 Rf1+ 26.Rf3 Re1 27.Kf5 Kd7 28.Rd3+ Ke8 29.Rd4 the game transposes to the line B) after 18.Rd4.
17.Ra4? lets the win slip: 17…Kd7 18.Ke2 Rb1 19.Ra6 Rb3 20.Kf2 Rb5 21.e6+ fxe6 22.Ra8 Rb2+ 23.Kg3 Rb3+ 24.Kh4 Rf3! 25.g5 hxg5+ 26.Kxg5 e5! 27.Kg6 Rg3+ 28.Kf5 Rf3+ 29.Kxe5 Rf1=.

15...Re1+ 16.Kd3 Kd8

16...Kd7 17.Ra7+ transposes into the main line. 

17.Ra8+ Kd7 18.Ra7+ Ke8

18...Ke6 19.Re7+ Kd5 20.Rd7+ Kc6 21.Rxf7 Rxe5 22.Rf8+-.

D. 1.-8



19.Re7+

This check is necessary for restricting the mobility of the Black king.

19...Kf8

Now White has to improve the position of his rook without giving the Black king the opportunity to come out from f8. 

20.Kd4 Rd1+ 21.Kc5 Rc1+ 22.Kd6 Rd1+ 23.Kc7 Ra1 24.Kd8

D. 1.-9



a) 24…Ra8+ 25.Kd7 Ra7+ 26.Kd6 Ra6+ 27.Kc5 Ra5+ 28.Kb6 Ra1 29.Ra7

White gains a decisive tempo. Now Black can defend only by transferring his rook to the back rank.

29…Rb1+ 30.Kc5 Rc1+ 31.Kd4 Rd1+ 32.Kc4 Rd8

Just continuing to check would not put more resistance: 32...Rc1+ 33.Kd3 Rd1+ 34.Ke2 Rd8 35.Kf3 Ke8 36.Ra1 Kf8 37.Kf4 Re8 38.g5 hxg5+ 39.Kxg5 Rxe5+ 40.Kh6 Ke8 41.Ra8+ Kd7 42.Kg7+-.

33.Kb5 Ke8 34.Kb6 Rd1

34...Kf8 35.Kc7 Ke8 36.Ra1 Rd7+ 37.Kc6 Rd8 38.g5 Rc8+ (38...h5 39.g6+-) 39.Kd6 Rd8+ 40.Kc7 Rd7+ 41.Kc6 Rd8 42.gxh6+-.

35.Ra8+ Kd7 36.Rf8

White rook finally reached this square.

36…Rb1+ 37.Kc5 Rc1+ 38.Kd4 Rd1+ 39.Kc3 Rc1+

39...Ke6 40.Re8+ Kd7 41.Re7++-.

40.Kd2 Rc4 41.Rxf7+ Ke8 42.Re7+ Kf8 43.Ke3 Rxg4 44.Ra7 Rg1 45.Ke4 Rf1 46.Kd5 Rb1 47.Ke6 Rb8 48.Kf5 Rb5 49.Rh7+-.

Back to D. 1.-9

b) 24...Rd1+ 25.Rd7 Rg1 26.Rd4 Rb1 27.Kd7 Ra1 28.Rd5 Re1 29.Ra5 Rd1+ 30.Kc6 Rc1+ 31.Kd5 Rd1+ 32.Ke4 Re1+ 33.Kd4 Rd1+ 34.Ke3 Ke8 

34...Rd8 35.Kf4 Kg8 36.g5 Rd1 37.g6 fxg6 38.Ra8+ Kf7 39.Ra7+ Kf8 40.e6 Rf1+ 41.Ke5 Re1+ 42.Kd6 Rd1+ 43.Kc7 Kg8 44.Ra8+ Kh7 45.f7+-.

35.Ra8+ Kd7 36.Rf8 Re1+

36...Ke6 37.Re8+ Kd7 38.Re7++-.

37.Kd4 Rd1+ 38.Kc3 Rc1+ 39.Kd2 Rc4 40.Rxf7++-.

Back to D. 1.-5 after 6...Ke8

B) 2.Rb4

Kopayev analysed only this move.

7…Rd1+ 8.Kc5 Rc1+ 9.Kd4?!

9.Rc4 is the best here. After 9...Rg1 10.Kd6 Rd1+ 11.Kc6 Rg1 we transpose to the main line A), D. 1.-6 after 8…Rg1.

9...Rd1+

a) 10.Kc4?!

Retreating to c5 wins as described above. 

10...Rc1+ 11.Kd3?! Kd7.

Kopayev stopped his analysis here and evaluated the position as drawn. In fact White’s position is still winning.

D. 1.-10



12.Ke4 Kc6

12...Re1+ 13.Kd5 Rd1+ 14.Rd4 Re1 15.Rd2 Ra1 16.Ke4+ Ke8 17.Rd4 Re1+ 18.Kf5+- transposes to the main line after 18.Rd4. 

13.Rd4 Re1+ 14.Kf5 Rf1+ 15.Rf4 Re1 16.Rf2 Kd7 17.Rd2+ Ke8 18.Rd4 Rf1+ 19.Rf4 Ra1 20.Rb4 Rf1+ 21.Ke4 Re1+ 22.Kd5 Rd1+ 23.Kc5

Now we are back to the position after 8.Kc5. 

23...Rc1+ 24.Rc4+-. Transposition to the main line after 9.Rc4.

b) 10.Kc3?! Re1

D. 1.-11



11.Rb8+?

This move of Kopayev throws away the win. White’s only move is 11.Kd4! Rd1+ 12.Kc5 Rc1+ 13.Rc4 with transposition to the above line.
11.Rb5? is also a result-changing error because of 11…Re4 or 11...Kd7=. 

11...Kd7 12.Rb7+ Ke6 13.Re7+ Kd5 14.Rxf7 14.Rxf7 Re3+?

After this move of Kopayev White again wins. Black draws quite simply with 14...Rxe5! 15.Re7 Re6!=.

15.Kd2! Rxe5 16.Re7?

Another error in the analysis of Kopayev. White wins easily after 16.Rd7+! Kc6 17.f7+-. 

16...Re6=.