- A non-trivial case where the side with the pawn is winning.
- A case in which the rook is between the pawn and its queening square and the attacking king on the other side of the pawn.
- The special case of the rook pawn.
First Case: Where the side with the pawn wins.
In the following case, the side with the pawn wins:
Fig 1. White to Play and Win
Note that both 1.Kf7? Rh7 and 1. Kg5? Rh1 threatening Rg1 will allow black to take the pawn and draw. Instead, white must play 1. Kf5 and walk up the board until black's rook runs out of room. Thus. 1.Kf5 Rh5+ 2. Kf4 Rh4 3. Kf3 Rh3 4.Kg2 and black has no threats left. Winning requires a bit more finesse in the following position, which is in all of the endgame books.
Fig 2. White to Move Wins!
Since this position does not show up once in the nearly 6 million games in my mega database, the practical player need not know it, but if you are interested the solution is here. The position in fig 1 is, by contrast, important to know as it is the building block to understanding more complicated positions. We will see this when looking at endgames in which a rook is fighting two pawns.
Second Case: Where the rook is between the pawn and its queening square and the attacking king on the other side of the pawn.
The following position is discussed in Yasser Seirawan's Winning Chess Endings (2003). White's goal is to win a tempo in such a way as to allow his king to get into play by stopping black from shouldering it out.
Fig 3. White to Move and Win!
Here white plays 1.Rd2+. If 1...Kd5 then 2.Rf2 Ke5, which gains enough time to win. If 1...Ke3 then 2.Rd1 threatening Re1+ winning the pawn. Thus black plays 2...f5 and again white will get his king into the game in time to win. For full analysis click here.
Third Case. The Rook Pawn.
I don't think the Rook versus single rook pawn ending is sufficiently explained in endgame manuals. As far as I know, the issues I discuss below are not mentioned in the chess literature, and the analysis I offer here is original.
There are two subcases.
- Where the rook attacks from the side along a rank and the king comes in from "above."
- Where the rook attacks from "above" along a file and the king comes in from the side.
In part one of this study of rook versus pawn endings we examined the following position (fig 4, the analysis of which is here) where black must play Kg5 to draw.
Fig 4. Black to Play and Draw!
1...Kg3?, by contrast loses here because of 2.Kf5 h3 3.Rb3+ Kg2 4.Kg4 h2 5.Rb2+ Kg1 6.Kg3 when we reach the position in fig 5:
Fig 5. Black to Play and Lose.
Black loses because 6...h1=N+ 7.Kf3 will result in the loss of the knight. So the key to saving this position for black is to prevent white's king from getting to g3.
Note that all of this works out differently with a non-rook pawn. In the position in fig 6 1...Kf3 (as well as 1...Kf5) draws.
Fig 6. Black to Play and Draw.
After 1...Kf3 2.Ke5 g3 3.Rb3+ Kf2 4.Kf4 g2 5. Rb2+ Kf1 6.Kf3 we reach the following position (fig 7), which is a draw because the knight can't be trapped with best play after 6...Ng1+.
Fig 7. Black to Play and Draw.
From this comparison of rook and non-rook pawns we can conclude that in rook versus rook pawn endings g3 is the key square for white. (Rather, to be more precise g3 is the key square for white when fighting against an h pawn, b3 is the key square for white when fighting against an a pawn, b6 is the key square for black when fighting against an a pawn, and g6 is the key square for black when fighting against an h pawn.) When attacking side can occupy the key square with his king and his rook is attacking from the side, he will win. So, the struggle in these endings is all about this key square. When the king is approaching this key square from "above" the attacking side must normally keep the rook attacking along a rank from the side.
This ending is often misplayed. Here we will look at common mistakes in recent play.
The following position was reached in 2012 in Marc Cluyts (1942) - Wim Barbier (2148) 0.5-0.5.
Fig 8. White to Play and Win!
White played 56.Rd8?, which throws away the win. The idea that "rooks belong behind passed pawns" is simply wrong here. White needs to check from the side to force the black king down the board and create room for the white king to move forward and eventually occupy g3. Thus, the main line 56.Rd4+ Kg3 57.Kf5 h3 58. Rd3+ Kg2 59.Kg4 h2 60.Rd2+ Kg1 61.Kg3, which reaches fig 5 above, winning for white.
In this case we have effectively the same position, but with colors reversed. Here the attacking side found a different way to throw away the win in Juergen Fritsch (1950) - Roger S Scowen (1883) 0.5-0.5 played in 2014.
Fig 9. Black to Play and Win!
In this case black played 53...Kd5? when 53...Rg5+ 54.Kb6 Kc4 etc. was needed.
The following, from a game in 2013, is a bit trickier. In the following position white's rook is attacked and cannot safely check the black king from the side.
Fig 10. White to Play and Win!
White played 67.Rf8?, which only draws. Instead, calmly withdrawing the rook from danger along the second rank wins: 67. Ra2 h3 68.Ra3+ Kg2 69. Kg4 h2 70.Ra2+ Kg1 71. Kg3 and white occupies the winning key square.
Not impressed by the errors made by these relatively low rated players? The following position, in which the defending side is a tempo ahead of the positions given in fig 8 and fig 9 above, appeared this year in Pia Cramling (2525) - Pontus Sjodahl (2420) 0-1.
Fig 11. White to Play and Draw!
Here GM Cramling threw away the draw with 88.Kc6? Perhaps she was thinking that she need to shoulder away the black king, but this move just wastes a tempo after 88...Kb4 89.Kb6 (what else?), and after 89...Re6+ 90.Kb7 Kb5 we see a familiar pattern whereby the defending side loses. By contrast, 88.Kb6 Re6+ 89.Kb7 Kc5 90.a7 Re7+ 91. Ka6! draws. If Cramling had focused on keeping black's king out of the b6 square she would have drawn. You don't have to be a GM to know this--at least now that you have read this blog.
Subcase 2. Where the Rook Attacks Along a File.
In the following case white's winning play is simple: advance the king and then--as per our prior blog posting--play Kg3 when the black King is on h1 and pawn on h2 to prevent stalemate and eventually win the h pawn.
Fig 12. White to Play and Win!
After 1.Kd6 h3 2.Ke5 Kh1 3.Kf4 h2 4.Kg3, we reach our known winning position. Knowing that the attacking side wins in the position in which the king is trapped in front of his rook pawn if it can play Kg3 helps solve other, more complicated problems.
In this case the attacking side needs to push the defending king in front of his rook pawn.
Fig 13. White to Move and Win!
In this position from an old game, Ernesto Rotunno - Neptali Ponce (Buenos Aires, 1939) 1-0, white has only one move that wins and it is 1.Rh1, the idea being that white will force the black king in front of the pawn after 1...h3. After this, play continued 2.Rg1+ Kh2 3. Rg8 Kh1 4.Kf2 h2 5.Kg3, reaching the winning position. The same idea appeared in 2003 in Hikaru Nakamura (2561) - Benen Samson (2251) 0-1. Can you beat Nakamura in the following position?
Fig 14. Black to Play and Beat Nakamura!
Samson correctly played 62...Rh8 [only move] 63.h5 Ke5 64.Kg6 Ke6 65.h6, Rg8+, forcing white's king in front of his h pawn, and reeled in the point.
The same basic idea applies here.
Fig 14. Black to Move and Win!
The only move that wins is 1...Rb1+, which was played in Vugar Ural Oglu Rasulov (2457) - Alexander Chudinovskih (2314) (Rijeka, 2010) 0-1. White has two choices: putting his king in front of the pawn, which loses for the reasons mentioned in our prior example, and 2.Kc6, which loses after 2...Kc8 3.a7 Ra1 4.Kb6 Ra2 after which white is in Zugzwang and loses the pawn.
Note that in this case and the preceding ones when the attacking king is on the side the rook must continue to attack along a file (ultimately the neighboring knight file) and not attack along a rank.
Here we consider the defense of such positions.
Fig 15. Black to Play and Draw!
Next: Rook against two or more pawns.