Tuesday, December 9, 2014

Maurice Ashley's thanksgiving study

The "Thanksgiving" is a festival celebrated in United States of America every year on the fourth Thursday of the month of November. People thank each other for the various things that they are grateful for in their life. GM Maurice Ashley, who is known to think in offbeat and unorthodox fashion decided to thank the chess world by composing a beautiful little study.

A small google search on Maurice shows that he is not only a chess grandmaster but also a commentator, author, app designer, puzzle inventor and motivational speaker! Truly a multi-faceted personality. And wait I almost forgot he is now also an organizer and the highest prize fund open tournament in the world, "Millionaire Chess tournament", was his brainchild.

Let's have a look at the composition!

White to play and draw.
My suggestion would be to take 15 minutes on the clock and write down your answer. After attempting it, have a look at the solution given below.

The first thing that we see in the position is that the black pawn is about to queen. But we must always starts our calculations with a check or a capture. Taking the knight is futile as it met with d1=Q. Giving a check to the Black king with 1.Rc3+ is a little more interesting but after 1...Kb6 2. Rb3+ Kc5 3.Rb1 Kc4 4.Rd1 Kc3 The d-pawn is going to cost an entire rook after which the win is trivial. After having convinced ourselves that no checks or captures work we settle down to the only move for White and that is 1.Rd3. The d2 pawn is now attacked. But it is easy to see that black has the very nice retort with 1...Nf3! 

With this tricky little move, the black knight boxes in the white king. Kg2 looks impossible because of the fork on e1 and if the rook moves to d5 with 2.Rd5 then after Kc6 3.Rd8 Kc5 4.Kg2 Nd4! The view of the rook is blocked and the pawn is promoted.

We now understand that White is in a huge spot of bother. Black king is marching down the board and there is nothing much that he can do about it.
But there is a way out of this mess!

Theoretical knowledge

Before we move ahead I would like to discuss two theoretical endgames with you.

King vs knight + pawn on seventh rank
It is well known that a knight and a rook pawn win against the opponent's lone king. (Unlike the bishops there is nothing like a wrong coloured knight!) But there is one drawn position which you have to be aware of and that is the following one:

Place the black king anywhere on the board, it doesn't really matter. All that white has to do is shuffle his king between a1 and b2 to make a draw. If black king comes too near on squares like a3, c3, c2 or c1 then it would be a stalemate.

Thus, we come to the conclusion that the knight that supports a rook pawn on seventh rank is a drawn position if the opposite king can stop the pawn from queening. A rook pawn on any other rank supported by knight from behind it will be a win. Hence, this is a unique draw.

Let's see the second theoretically drawn position.

This position is also drawn irrespective of who is to move.

The white king keeps moving between b1 and a1 and there is no way to win. Once again, if the black king tries to come closer, it will end in a stalemate.

Knowledge of these two theoretically drawn positions is crucial for solving the Ashley study. If you do not know them you will be like Mr. Houdini who thinks that Black is winning. Don't believe me? Here is the proof:

As you can see on the lower right screen of the corner, Houdini 4 thinks that this position is won for black. It's time to prove it wrong! :)

We can see the contours of the two theoretical drawn positions over here. All that we have to do is to eliminate the d2 pawn. How would you continue?

There are two different ways to get rid of the little guy on d2. One is playing Kg2 and the other is taking the pawn directly with Rxd2.
Which variation shall we begin with? I always prefer to start with captures and forcing variations and hence 2.Rxd2 is the move that I would analyze first. After Rxd2 Nxd2 3.Kg2 we reach the following position.

We already know that if the white king were to settle down on the b1 square it would be a draw. He is still quite a few moves away but we need to be accurate as Black. How should black play?

I personally find such positions extremely difficult to calculate, mainly because there are so many plans and possibilities available for both sides. Let us start with the most natural plan of capturing the a2 pawn. After 3...Nb1 4.Kf3 Nc3 5.Ke3 Nxa2 6.Kd3 Nb4 7.Kc3 a2 8. Kb2, the white king arrives in the nick of time to stop the pawn.

We already know that this is a theoretically drawn position.

So is this position a draw?

No, it is not. There is a win in there which is not so easy to find but strong technical players like Karpov and Kramnik would solve this position very easily. And how exactly do they do it?
They think of the positions that their opponent would want to reach and try to stop their idea. Long term prophylaxis you can say!
In the above position, white would like to get his king to b1 to reach a theoretically drawn position. Now is the difficult part. We have to put all our energy as black here and try to stop the idea in the best possible way. I am sure if you give your 100% you will be able to find the winning setup of putting our knight on c3 to prevent the king from coming to b1 and the black king on b4 to protect the knight.

Putting these ideas into moves we find 3...Kb6 4.Kf2 Kb5 5.Ke2 Nb1 6.Kd3 Kb4 7.Kc2 Nc3! -+ 

And there you have it! The white king has been prevented from going to b1 and next black takes the pawn on a2 and doesn't make the mistake of pushing his own pawn to a2. He will first bring his king to b3 and then play a3-a2 to earn a well deserved full point.

We now understand that 2. Rxd2 doesn't quite work. What is now left is the second option with 2.Kg2!


Black goes 2...Ne1+ 3.Kf2 Nxd3 4.Ke2

White wins the d2 pawn and in much more favourable conditions than in the last variation. His king is much closer to the queenside, especially the b1 square.

The only option left for Black now is to win the a2 pawn as bringing his king would be too slow. So after 4...Nb4 5.Kxd2 Nxa2 6.Kc2 Nb4 7.Kb3 a2 8.Kb2=

It's a draw! I have shown you this position so many times that you are not going to forget it for the rest of your life.

As you could see that in order to solve the study successfully you not only needed the theoretical knowledge of two positions that we saw but also hardcore calculating and execution skills in order to decide between 2. Rxd2 and 2.Kg2. As it turned out the latter was the right way to start.
Congratulations if you were able to solve this position correctly. We must definitely thank Maurice Ashley for composing this beautiful study.

By solving this position I came to a very important conclusion which I would like to share with all of you:
There are two very important things that a chess player must possess in order to improve.
1. Sound theoretical knowledge (opening, middlegame and endgame)
2. Strong execution (calculating) skills.

Lack of theoretical knowledge with good calculating skills is like going in a car without headlights on a dark road. You will keep moving but you don't know whether you are going in the right direction. Not to mention, a huge accident might just be waiting to happen!

At the same time good knowledge without execution skills (calculating ability) is like going in a car that has run out of fuel. You can see where you want to go but just don't have the resources to reach there! :)

Improve both the above mentioned skills and you will zoom ahead on the road of chess improvement!


  1. Keep up with the good work!
    I'm quite disappointed that Houdini Four thinks that that position is won for black. How to trust any commercial engine after seeing this??

  2. Most commercial engines assume that tablebases are installed, so they don't worry about misevaluating positions with six or fewer pieces on the board. They can still get fooled in more complicated positions, of course.

  3. This comment has been removed by a blog administrator.

  4. i wondered why Rc3+ and then Rd3 would lose and I found a cool line. After 1. Rc3+ Kb6 2. Rd3 Nf3 3. Kg2 Ne1+ 4. Kf2 Nxd3 5. Ke2 Nb4 6. Kxd2 and Nxa2 is a draw so black plays Kb5! (or Ka5)so that after 7. Kc3 there is Ka4, preventing Kb3. If 7. Kc1 then Nxa2+ and the king is close enough to prevent white from forcing a2.