The secret of As-Souli

By: Yuri Averbakh

In Istanbul there is an ancient chess manuscript. It was written in 1140.

This manuscript contains about 200 positions. One is shown below:


What rubbish, you think? 

Queen and King against a lone King. Any beginner can give checkmate in this position.

Don't be so hasty! You must remember that in ancient times chess was not the same as it is today.

In shatranj (as the game was called) the Queen could move only to the four nearest squares on the diagonals, backwards and forwards.

The Farzin (the old name for the queen) was a weak piece and it was impossible to checkmate a lone king with a king and a farzin.

But it was possible to win if you could kill all the opponent's pieces. This situation in shatranj had a special name - "bare king". If the king was left alone, the game was over. This meant that the position above could not happen in shatranj. Something was wrong.

You may ask a question. What was so interesting in this position?

The answer - the text that was connected to it.

This position is very old and extremely difficult to solve. Nobody, not even Al-AdIi, could solve it, or say it is a draw or that white wins.

"There is no man on earth who can solve it if I have not shown him the solution."

These are the words of As-Souli as added in the manuscript.

As-Souli! Historians know this name well. And not only historians. Abu-Bakr-Mohammad Bin Jahia As-Souli was born at the end of 9th century.

He lived in Baghdad and died in 946 in Basra. By ancestry a prince from Middle Asia, he was statesmen, poet, literature expert and historian. His works about poets of the Islamic period and about the history of Abassid's house survive and are well known today.

Besides this As-Souli was a great chess player. He wrote two books on chess but these have not survived. We know about them from other sources.

In his time he was a real champion, at least a champion of Arabian countries.

What do you think about the words of As-Souli? He is boasting, is he not?

When I discovered that in the As-Souli position there were some errors I was greatly disappointed because I badly wanted to solve it. Then I was lucky enough to remember a similar position: 


White wins. The problem is how to kill the opponent's firzan to reach the "bare king" position.

This position was analysed by Al-Adli. You remember that As-Souli speaks about Al-Adli. Al-Adli lived in Baghdad much earlier than As-Souli. He was the Khaliph's court chess player and is counted as the first author of chess works written in Arabic. His works have also not survived, but we have extracts from many other manuscripts.    

The manuscript's authors say the position in the second diagram belongs to Al-Adli. He gives these moves to win:

1.Ke6 Kf4 2.Kf6 Kg4 3.Kg6 Kh4 4.Qg5+ (after 4.Kh7 Kh5 draw) 4...Kg4 5.Qf6 Kf4 6.Kf7 Kf5 7.Qe7 Ke5 8.Kg8 Ke6 9 Qf8 wins.

As I have already said, As-Souli lived after Al-Adli and so checked all his positions.

He found an elegant and shorter solution:

1.Kf8! Kf5 2.Kf7! Kg4 3.Kg8 Kh5 4.Kh7!

The reader can see that it is a zugzwang position. The black king must retreat and white wins.

Are you not astonished? It means that zugzwang (originally a German word used by problemists only at the end of the 19th century) was well known a thousand years ago.

Because of Al-Adli's position, I started to think that the As-Souli position may be similar with kings and queens only. But where should I put the black firzan? It seems that the firzan should be on a1.


It is clear that white should try to win. But who is to move?

To find this was easy. In the manuscript it says: if white to move, he wins in three moves. It means that black is to move.

By the way, we can check if our position is correct. We must show that white wins in just three moves. Let's prove it:

  1. Ka2 Kd3 2.Qb4 Kc4 3.Qa3.

We are right, the position is correct. We have no contradictions.

Much later I found a similar position in a manuscript of Abulfath (11th-12th century) which is in Tashkent (USSR).

Now I could try to solve As-Souli's position. Something was said about the solution in the Istanbul manuscript:

"Black has no moves excluding to move his shah (king) on a square of his firzan which is a knight's move distance from the red firzan. If he moves to any other square he loses. Then the red shah goes on the fourth square of his firzan.”

The author of the manuscript makes the following remark: "Now the solution ends, partly because it is very long, partly because As-Souli was very proud of it."

By the way, you can see that in the Arabian manuscripts pieces were not white and black as today, but red and black because red and black ink was used.

First of all I translated the moves from the manuscript into our algebraic notation.

It was easy.

1...Kd5 2.Kb4 Kd6. 

Now we must prove the words of As-Souli; that 1...Kd5 is only move.

If 1...Kd3 then 2.Qb4 and 3. Ka2 wins. 

If any other move, then 2.Ka2. Everything is okay!

Now we must find the right answer to 2...Kd6.


Let's try 3.Qd2 Kd5 4.Kc3 Ke4 5.Kc2


At first I decided that the problem was solved, but after a while I found the only move 5...Kf3!. 

If 6.Kb1 then 6...Kc2 7.Qc1 Qd1 and it is white who is in zugzwang.

Then I tried again to advance the king.

3.Kc4 Ke5


4.Qb4 Kd6!

I do not see any other answers. After 4...Ke6 or 4...Ke4 5.Kb3 Kd5 6.Ka2 Kc4 white wins.

5.Kc3! Kc6

If 5...Kd5 then 6. Kc2 Kc4 and the rest is well known already: 7.Qa3 Kb5 8. Kb1 Ka4 9.Ka2 etc.

6. Kb3 Kb5 7.Qc3 Kc5 8. Kc2 Kc4 9.Qd2 and white wins as was already shown. 

It looks as if I had found a solution. But simultaneously I had a feeling that something in this solution was wrong. I had my doubts that such an expert of chess as As-Souli could praise this solution so highly.

I decided to check my solution once more.

Why, for instance, after 3.Kc4 should black play 3...Ke5? I tried 3..Ke6


After 4.Qb4 I found a fine move; 4...Kd7!! 

If 5. Kb3, then black plays 5...Kc6! 6.Ka2 Kb5 with a draw. And if 6. Kc3 then 5...Kd6! with the same result.

It means that my first try was unsuccessful.

I then decided to attack the position of As-Souli’s problem using the invention of our century - the method of corresponding squares.


It is clear that the kings here are maneuvering  on corresponding squares. d7 corresponds to c4; d6-c3; c6-b3. 

We can mark these squares by letters. Playing 4.Qb4 white cannot win the battle for corresponding squares. I decided to advance the white king. 

4.Kd4 Kf6


After I put the black king on f6, my memory told me that somewhere I had seen a similar symmetrical position before. I checked the Istanbul manuscript and discovered this position:


White wins. The solution of this position is 1.Kg6 Kg8 2.Qd2! Kf8 3.Qc1

Now the kings go back but the white king wins the race.

3...Kc7 4.Kf5 Kd6 5.Kc4 Kc5 6.Kd3 Kb4 7.Kc2 Ka3 8.Kb1 and the black queen is lost.

Then it was like a thunderbolt in the sky. I got the idea of trying to transfer the position of As-Suoli in the last one. It can be done very easily.

White: Kd4, Qc3

Black: Kf6, Qa1

I immediately took my pen and started to write:

5.Kd5 Kf7 6.Ke5 Kg7 7.Ke6 Kg8 8.Kf6 Kh8. Bravo!

It was as if somebody else moved my pen.

Now having solved As-Souli's problem, it was possible to judge it properly. After the black king is pressed to come to the h8 square playing 9.Kg6! white really wins the battle for the corresponding squares. The chess board in this case is getting too small! The corresponding square to g6 is i9 but it is out of the board.

And what about the solution itself? Is it not amazing? Both kings are running from one corner to another and then back. It is a creation of genius. I do not know any modem endgame study with such an idea. 

Honestly, when I first read the words of As-Souli, I thought he was just boasting. Now I can say that he had an absolute right to say such words.


This is