Wednesday, September 19, 2007

An exercise in backwards thinking

The same position as before but now solved with backwards thinking.

View from defender
You start with identifying the invasion squares. That are the targets you intend to attack. a2 is already under atack by the queen, e4 is under attack by the B, R and N. Look at the defenders of those targets too.
If you think backwards, you imagine the previous targets c3 and c5:

View from attacker.

On c3 stands as target an overloaded knight with such important function (protecting a2) that is has a high value, on c5 the queen with a high value too. So that are natural targets for a black knight on e4. A black knight would be taken off the board immediately by the white knight on g3, which has its target square on e4.

View from defender.

This shows that the knight on g3 is the piece that spoils the party. So you have to look for ways to deflect that knight first.

View from attacker.

If you take with the rook on e4 first you deflect the knight on g3 since you have some serious threats. The targets of the discovered attack are the white king and queen. So white must take with knight g3.

So the main line becomes 1. ... Rxe4 2. Nxe4 Nxe4 3. Nxe4 Qxa2+ leading to mate.
Ok, this story is somewhat feeble and shaky, and not quite waterproof, but I hope you get the idea. The point is to recognize the characteristics, which means identifying the targets, and work your way back to the beginning. I will have to do a lot more exercises with backwards thinking to get experience. The great power of the method is that it rules out moves that aren't aiming towards targets. This prunes the tree drastically, thus preventing brain damage.

This is a typical alpha, bhèta, gamma type of forced attack, where alpha= prelimanary skirmishes, bhèta=invasion, gamma=deliver mate.

For backwards thinking in relation to accidental tactics you have to start with the recognition of pieces of high value as potential targets. Accidental tactics aren't about squares. The value of a piece can be high because it's static value is high or because it's dynamic value is high (e.g. it fulfills an inmportant function, like white's overloaded knight on c3 above)


  1. As I recall, I believe that you came up with simalar conclusion in your endgame study of a year or so ago. It was helpful to begin with an end in mind and work backwards.

    Doing studies like this emphasize the importance of knowing basic mating paterns/piece winning patterns. The ability to explore a win and work backwards to see how one could get there.

    I am finding that these complex positions are at times at a crossroads of 2 or 3 thematic attacks/mates. The complexity lies in the shifting the pieces around to see if you can reach it. In many cases, the treat of reaching one type of mate or material advantage, opens the door for the other attack.

    The trick is finding these forcing moves that are advantageous to you.

  2. Tak,
    if I recall well you were even the one who suggested to start at the end. In middlegame positions the end is too far away and hidden. But to start at position bhèta or gamma is doable fairly well, since most people don't go beyond trial, error and gambling.

    The difficulty lies in the invasion squares. They are much harder to see as targets than pieces of high value. But the same simple tactical motifs as we use to gain wood can be applied to gain squares.

  3. Good example of a tactic that requires preparation. I think most tactics do, unless your opponent is so kind to blunder. Stepping backwards is obviously the action to be taken. I think you got the point: First you must identify the goal and then step backwards to see how to reach it, instead of using trial and error forwards from the actual position.