When in Doubt, Play it Out Print E-mail
By GM Joel Benjamin   
October 8, 2007
Dear Joel,

I was playing in a simul at a local library against a master, rated 2267. My rating is 1203. In an unconventional manner he played Black on every board.  Results: 1 loss to Brazil's National Champion, 1 draw. He won the other 10-12 games.

My game was the draw.  Here is the final position:


Here are the last few moves of the game:

45. Nb4+  Kc5
46. Nd3+  Kd5
47. Nb4+  Kc5
48. Nd3+  Kd5
49. Nb4+
Draw by repetition

I have heard many people in the game say, "When in doubt push a pawn".  I have learned from experience this is not correct.  I prefer when in doubt simplify the position.  I remember a game once where this helped me win the game.

Here is my question:  I was looking for a good simplification and could not
see it.   I also had not played in couple of years.  I decided to play for draw because I could not find a win or simplify to my satisfaction.  Also my opponent was rated 1000 points above me.  Is this a good strategy in this type of a situation?

Computer analysis does find a win for white but appears to require precise play.



I’m not sure how I would finish “when in doubt…”—perhaps this would make a good top ten list for Jennifer.  I do know that “push a pawn” is certainly not the answer.  Pawns can never retreat, and ill-advised pawn moves can be especially costly in the endgame.

It is so tempting to take a draw against a much higher-rated player.  Very often when such an opportunity arises, we start to put our standards unreasonably high.  If we don’t see the “win” we take the draw.  Here White is a pawn up with the potential of forming majorities on both sides.  With limited forces, Black’s winning chances are extremely low.  You would have nothing to lose by playing this out, even if you weren’t sure how you should continue.

The most obvious continuation is 45.h4, looking to create a dangerous outside passed h-pawn.  This would win easily if not for the tricky response 45…f4+!

Position after 45..f4!

Now 46.gxf4 gxh4 turns the tables, while 47.Kf2 fxg3+ 48.Kxg3 gxh4+ allows Black to simplify and draw easily.  [Remember, when ahead in the endgame, trade pieces, not pawns] 

White has a good winning chance in 46.Nxf4! gxf4+ 47.Kxf4. 
Position after 47.Kxf4

After 47…Ne2+ 48.Kg4 Nc3 49.Kf5! Kd6 50.Kg6 Ke7 51.Kg7 Black cannot grab the a4-pawn without the h-pawn wreaking havoc.  With his king unable to help, Black can sit around waiting for a pawn to queen, or fall short in the line 51…Nxa4 52.h5 Nc5 53.h6 Ne6+ 54.Kg8 Ng5 55.h7 Nxh7 56.Kxh7 Kf6 57.Kh6 Kf5 58.Kh5.  White will queen the g-pawn.

Admittedly, this line is not obvious at your level, though you should have played 45.h4 anyway if you didn’t see 45…f4+!

Following your train of thought regarding simplification, 45.Nb4+ Kc5 46.Nc6! would have fit the bill nicely.  I assume you rejected it because you lose a pawn after 46…Nxc6 47.bxc6 Kxc6, but then 47.h4 gives you a basic outside passed pawn win.  47…gxh4 (47…f4+ 48.Kf3! only makes matters worse for Black) 48.gxh4 Kd5 49.Kf4 Ke6 50.Kg5!

Position after 50.Kg5!

Black can’t “race” with 50…Ke5 51.h5 f4 52.h6 because White will queen with check.  Alternatively, 50…a5 51.h5 f4 52.Kxf4 Kf6 53.Ke4 Kg5 54.Kd5 Kxh5 55.Kc5 Kg6 56.Kb6 Kf7 57.Kxa5 Ke7 58.Kb6 Kd7 59.Kb7 leads to a White queen on the other side.  This line may look long and complicated, but there are very few branches along the way.  Even if you didn’t see it to the end, the most you risk is a draw anyway.

Black could decline the trade with 46…Ne6, but 47.Nxa7 puts you two pawns up.  Black can struggle but not save the game.  The following line suggests one possible win; though it isn’t forced, it does illustrate the problems knights have with outside passers:  47…Kb6 48.Nc8+ Ka5 49.Nd6 f4+ 50..Kf3 Nd4+ 51.Kg4 fxg3 52.Kxg3 Kxa4 53.b6 Kb4 54.b7 Nc6 55.Kg4 Kc5 56.Nf7 Kb6

Position after 56...Kb6

 57.Kxg5 Kxb7 58.h4 Ne7 59.h5 Ng8 60.Nh6 Ne7 61.Nf5 Ng8 62.Kg6 Kc6 63.Kg7 and after taking the trapped knight White queens the h-pawn.

These variations may seem involved but are actually rather straightforward.  Both winning methods start with logical plans.  Your biggest mistake was not having faith in your superior position leading you down the correct path.  So my advice to you might well be, “When in doubt, play it out.”

To ask GM Joel a question, read the submission guidelines , and email Joel at [email protected]