Many games that seem like dead draws are actually very winnable.
And strangely, one of your best chances to fight for a win can be to liquidate into a pawn endgame – even when you’re down a pawn!
Over the years some of my most memorable victories have come from pawn endgames in seemingly drawish positions.
You may have already read my article about the importance of learning endgames, but how do you know when you should exchange pieces and fight an endgame? And how should you try to calculate them?
My hope is that by showing you a few of my own games and calculations, you’ll be able to answer these questions yourself.
Without further ado, let’s start with a position from a game I played against Hrant Melkumyan back in 2009:
Naturally, the first thing I did was to consider the move 1.Rb6.
After 1...Rxb6 (It is important to bring the enemy’s pawn closer, because in a pawn endgame the further the distant passed pawn is from the zone of action, the more dangerous it is!) 2.axb6 Kd6 3.b7 Kc7 4.Kd5
(Be aware that 4.Ke5 doesn’t work, because of Kxb7 5.Kf6 Kb6 6.Kg7 Ka5 7.Kxh7 Kb4 8.Kxg6 Kxc4 and the pawns pass simultaneously).
The following position arises after white plays 4. Kd5:
After 4...Kxb7 5.Kxc5 I quickly established that Black cannot resist White's plan – sacrificing the c4 pawn in order to distract the enemy’s King by breaking through to the black pawns.
However, Black has a resource - 4...h5! (not g5? 5.g4!) 5.Kxc5 g5 6.Kd5 h4 7.Ke4 g4 8.Kf4 h3 and Black successfully exchanges the pawns on the kingside.
Having calculated this variation, I figured out that the pawn endgame is drawish and in the initial position played 1.Ra1?.
My opponent Hrant defended well and the game finished in a draw.
Only at home I found out that the pawn endgame was winning...
Yeah Black’s 4…h5 was a nice resource, but it was not saving Black as I thought during the game.
Unfortunately, during the game I missed that instead of 5.Kxc5 I can rush to the kingside pawns, as with the last move, Black’s pawns came closer to my king, like a fish getting closer to a shark.
Now, after White played 4...h5, there are 2 factors in White's favor:
1) The hunt for the pawns becomes shorter.
2) After capturing the h5-pawn, the White King does not block the advance of his own pawn.
I could play 5.Ke5 Kxb7 6.Kf6 Kb6 7.Kxg6 Ka5 8.Kxh5 Kb4 9.g4 Kxc4 10.g5 Kd3 11.g6 c4 12.g7 c3 13.g8Q c2
If you’ve watched the Queen vs Pawns course, you should already know this theoretical position. Because of the “c” pawn, we should be careful to avoid Black’s stalemate idea. Luckily, there is one move, which gives White victory...
Just in time! 14…Kc3 15.Qc1! and White wins.
In my calculations, I simply missed 5.Ke5! and couldn’t win this important game…
It always hurts to be one step away from victory.
However, I made the right conclusions and after the tournament, I began studying pawn endgames more carefully.
An opportunity to demonstrate what I learned was presented unexpectedly after 2 years against the same opponent – GM Melkumyan.
At the time, this game was of great importance. Hrant was the leader of the tournament and I was only one point behind. But this time, having faced my good friend again, I did not miss the opportunity to transition into a pawn endgame which led to victory.
Black wanted to play 1...Kh8!
So I didn’t have time for 1.Kg2, 2.Kf3 and then h8Q
Rxh8 2.Rxh8 Kxh8 3.Kg2
It’s incredible! After the exchange of Rooks, a symmetrical position is obtained with the only difference that Black has an extra pawn on f6! Why should White win?
The point is that White manages to occupy the key f4 square with his King, after which Black gets into zugzwang and is forced to retreat, losing the pawns.
3… Kg7 4.Kg3 f5
(If 4…Kg6 5.Kf4 Black is lost because of zugzwang. 5…Kg7 6.Kf5! after which I win the f6 pawn, and then the d5 through Ke5)
5. Kf4 Kf6 6.f3
Black got into zugzwang and is forced to let the enemy’s King pass on g5 or e5․ Although the final outcome is still in question, it is clear that it is White who is playing to win.
This means that in a practical game it is enough to calculate up to here and play h8Q, which I did in the game, as with Rooks on the board, there was no chance of winning.
Back to the game, now Black has two reasonable moves – 6…Kg6 and 6…Ke6.
In case of 6...Kg6 7.Ke5 Kg5 8.Kxd5 Kf4 7.Kc4 Kxf3 8.d5 Kg2 9.d6 f4 10.d7 f3 11.d8Q f2 the following position arises:
If there was not a pawn on f7, it would be a draw.
But now Black is lost because of the lack of a stalemate idea. 12.Qg5+ Kh2 13.Qf4+ Kg2 14.Qg4+ Kh2 15.Qf3 Kg1 16.Qg3+, and there is no stalemate after 16…Kh1 17.Qf2.
In the game, instead of 6...Kg6, Melkumyan played 6...Ke6 7. Kg5 f4
(Or 7...f6+ 8.Kf4! Ke7 9.Kxf5 Kf7 10.f4 and Black is again in zugzwang.)
Now if I play 8. Кхf4? Kf6 it would be a draw.
However, the thing is that after 8…Kf6, Black is in zugzwang if it was again his turn to play. So what should be played?
Now if 8…Kf6 9.Kxf4 and Black loses because of the Zugzwang.
8...Ke7 9. Kf5!
Finally displacing the opponent's King.
9...f6 10. Kxf4
Only now, when Black is not able to occupy the key f6 square with the King, White captures the pawn. The rest is simple
10...Ke6 11. Kg4 Ke7 12. Kh5 Kf7 13. Kh6 f5 14.f4
After realizing how active my King was, and that he’d lose his pawns due to zugzwang, Hrant resigned.
Inspired not only by my victory over the leader of the tournament and a formidable opponent, but also by the fact that my work on pawn endgames had not been done in vain, I had a few more victories and became the champion of Armenia!
Let’s move to another example.
A curious incident happened to me in the following game:
We had just passed move 40, and thirty minutes were added for both of us. My opponent pondered…
I clearly want to activate my Queen with 1...Qd6.
For Godena it makes sense to think about 1.Qe6, but then he should precisely calculate the pawn endgame which occurs after 1...Qd6 2.Qd6 cd6.
In this position I’m a pawn up, but the problem is that I can’t break through the queenside and if my opponent can close the kingside, I’ll not be able to win.
Now 3.g4? is clearly bad – Black will transfer the King to e6 and play f5.
White’s other option is 3.Kf2 Kc7 4. Kf3
Now I would love to play 4...f5, but then White has 5.g4!
If I take on g4 then I can’t defend the g5 pawn.
And if I play 5...f4 I wouldn’t be able to break through despite the fact that I’m up a whole connected passed pawn!
All that remains is 4...Kd7 5.Kg4 Ke6 6.b4 b6 7.b5!
White closes the quenside, and most importantly, now it’s a position of mutual zugzwang, and it’s my move.
I’m forced to let the White king pass.
7...Kе7 8.Kf5 Kf7 9.g4 Ke7
It's bad to play 9...Kg7? 10.Ke6 Kg6 11.Kxd5 f5 12.gxf5+ Kxf5 13.c4 with the idea c5, and White wins.
It seems that Black has achieved their goal, 11. Kg7 is followed by f5, and 11. Kh6 is followed by Kf7!, and White is in zugzwang.
However, White has 11. Kh5!! Kf7 12. Kh6, with a draw.
Having calculated this long variation, I looked at the clock - my opponent had 10 minutes left, and he continued thinking. I began to think about what I would do after 1.Qe6.
Another 5 minutes passed, and my opponent, already tired and having only 5 minutes left on the clock, finally played 1.Qe6, but after an instant c6! could no longer hold this endgame after being overwhelmed by long calculations and having little time left on the clock.
Something quite comical happened to me in the following game:
White has an extra pawn, but the position is clearly leading to a draw – like most Rook endgames when the pawns are on one side of the board.
I was thinking about playing 1... f6!?. But I suddenly saw the possibility of a devilish trap, into which my opponent unexpectedly fell!
1...Rb1!? 2. Rf1?? Rxf1 3. Kxf1 Kf6
It's incredible! Who would have thought that this pawn endgame could be lost for White?
Here my opponent played 4. Kf2 Ke6 5. Kg2 Kd5!
It is important to bypass the e5 square!
6. Kf3 Ke5 – White is in zugzwang.
7. Kg2 Ke4 8. Kf2 f6
Another zugzwang! And This is the whole point, I had to have this extra move to win the game.
9. Kg2 Kxe3 10. Kf1
It wasn't too late to stumble – 10...f4?? 11.g4! hg4 12.h5
11. Kg2 f4! 12.gxf4
If 12.g4 then after 12…hg4 13.h5 Kf5 I’m in time.
12…Kxf4 13. Kf2 Kg4 and soon White resigned.
Instead of 4.Kf2 White could play 4.Kg2 which would be a much better option for them.
4…Ke6 5. Kf2!
Now after 5...Kd5 White has an unexpected resource.
6. е4+! fxe4
Or 6...Kxe4 7. e3, and White successfully defends.
7.g4!! hxg4 8.h5 Ke5 9.h6 Kf6 10.Kg3! - taking advantage of the fact that the Black King has blocked the way of the f pawn, with a draw.
So was White really saved? It turns out...no!
Instead of 5...Kd5 The victory lies in the unexpected 5...Kd6!!
Now if 6.Kg2 Kd5 leads to a win as after 7.e4 Ke4 White can’t play 8.e3.
And if 6.е4 fxe4 7.g4 hg4 8.h5 Ke7! 9.h6 Kf8! 10.Kg3 f5 and Black wins.
I’m going to play 10...Kg8, 11...Kh7 and capture White’s pawn.
White can’t resist it, as they can’t play Kf4, Kg5 or Kf4, Kf5, because of the move g3.
Now that you’ve learned the importance of zugzwang in endgames, and seen a few examples of how to calculate pawn endgames, I have a challenge!
Below is a game I played against Boroljub Zlatanovic:
It’s White’s turn to move.
Does 1. Qd3+ lead to a win or not?
I look forward to seeing your answer in our forum.