HERE is a a fascinating and instructive endgame position from Carl Portman's Never Mind The Grandmasters column in Chess magazine.
 |
| Black to make his 33rd move in Portman (1888 ECF) - Tayeb Kazemi (1791 ECF), Leamington (Warwickshire) League 2026 |
*****
*****
*****
*****
*****
*****
Pawn endings with rival majorities are tricky, no matter how equal the engines reckon the position is (completely equal, according to Stockfish18 and Dragon1).
So how does one go about gaining winning chances, or, at least, setting the opponent problems?
The answer, I believe, is to try to create a passed pawn that will divert the enemy king, leaving your own king free to attack weak enemy pawns, and thus create another, this-time-winning, passer.
Here is an example from a 1991 rapid game at London's Barbican.
 |
| Black has just captured a knight on e6 in Spanton (151 BCF) - Martin Cutmore (152 BCF) |
*****
*****
*****
*****
*****
I played 26.Kf3, as advancing the king is rarely wrong in such positions, although the engines marginally prefer 26.g4!?
After 26...h5, I took the chance to also advance my backward pawn with 27.e4, and following 27...dxe4+ 28.Kxe4, we have a position with rival pawn-majorities.
 |
| The game is completely equal, according to Stockfish18 and Dragon1 |
*****
*****
*****
*****
*****
28...f5+?
This natural-looking move not only gains space, but comes with a check, and yet after it Black is lost. The engines reckon several moves maintain the balance, including 28...g6, 28...a5 and 28...b5.
The problem with the text is that after 29.Kf4, Black is obliged to defend the kingside as 29...Kd5 30.Kxf5 Kxd4 is hopeless, eg 31.Kg6 Ke3 32.Kxg7 Kf4 33.Kh6 Kg4 34.g3, when Black's king will eventually have to move, allowing Kxh5.
My opponent played 29...Kf6 (29...g6 runs into the same trouble as in the game after 30.Kg5 Kf7 31.c4), but 30.c4 g6 31.d5 cxd5 32.cxd5 created a diverting passed pawn.
The game finished 32...a5 33.b3 b5 34.a4 b4 35.g3 Ke7 36.Ke5 1-0.
Returning to Portman-Kazemi, what would you play in the diagram position?
*****
*****
*****
*****
*****
The game saw 33...h5?, which, probably to both players' surprise, loses.
CP states that, after the move, "the engines goes up to '+7' for White." I do not know which engine he is talking about, but for a long time Stockfish18 puts White's advantage at a little over 3.00, while Dragon1 gives White only the upper hand. However, after allowing them more time than I usually would, Stockfish18's evaluation rises to just over +4.00, while Dragon1 agrees White is winning, but puts the advantage at a little under +2.00.
Even so, whatever the precise number, there seems little doubt White is winning, whereas moves such as 33...g4!? (CP's suggestion) and 33...a5 keep the game level.
I might have been inclined to play 33...Kd5, and if, as in the game, 34.c4, a sample continuation runs 34...bxc4+ 35.bxc4+ Kd6, when neither side is able to safely create a diverting passer.
In the game, White met Black's faulty 33rd move by gaining space with 34.c4!, which the engines agree is the strongest continuation by some distance.
There followed 34...a6 35.a4 Kd6 36.a5 Ke6.
 |
| How should White proceed? |
*****
*****
*****
*****
*****
37.Ke3
Also winning is 27.h4 - the idea will become apparent as the game progresses.
However, simply creating a passed pawn is NOT the way to go. Indeed, 37.cxb5?? turns a win into a loss after 37...cxb5 as, far from diverting Black's king, the sequence allows the king to advance to the d5 square, eg 38.Ke3 Kd5 39.Kd3 f4 40.gxf4 gxf4, when it is Black who has created a diverting passer, a possible continuation being 41.b4 f3!? 42.Ke3 f2 43.Kxf2 Kxd4 etc, or 41.h4 b4 42.Ke2 Kxd4 etc.
37...Kd6 38.h4!?
This could also have been played in answer to 37...Kf6.
38...bxc4 39.bxc4 gxh4
Or 39...g4 40.Kf4 Ke6 41.Kg5 etc.
40.gxh4 Ke6
 |
| Black has created a passed pawn, but rather than diverting White's king from a crucial part of the board, the sequence has created weak squares that invite the king to advance |
*****
*****
*****
*****
*****
41.Kf4 Kf6 42.d5!?
But this does create a passer that diverts.
42...cxd5
Or 42...c5 43.d6!? Ke6 44.d7 Kxd7 45.Kxf5 etc.
43.cxd5 Ke7
As CP points out, 43...Kg6 loses to 44.Ke5.
44.Kxf5 (1-0, 53 moves).
So why was 33...h5? weak? The answer is it eventually let White create kingside weaknesses in Black's position with 38.h4!?
Could Black have prevented this? Well, he could have prevented the h4 thrust, eg with 34...g4, but that does White's job for him by immediately creating black kingside weaknesses, a likely continuation being 35.Ke3 a6 36.Kf4 Kf6 37.cxb5 cxb5 (37...axb5 38.a4) 38.d5, when the diverting passer wins.
As is often the case, a 'simple' pawn ending has proved to be anything but simple.
