Saturday, 28 August 2010

Injustice Positions

While flicking through a Russian language book on endings that I had in my library, I came across a set of positions. Even though I cannot read a word of Russian, I recognised the positions as 'Injustice patterns'. The 'injustice pattern' is a position where it looks as though one side 'must' be winning, but it turns out that the positions is instead drawn. One obvious example, indeed so obvious that the description is never applied to it, is K+B+rP v K, where the bishop does not cover the queening square, and the defending king does. A not so obvious example is given in the first diagram. Here White has the right coloured Bishop but the presence of the pawns on a6 and a7 mean the position is drawn any way (although a number of my computer engines spend a lot of time thinking it is winning for White!). You can even shift the position 1 file to the right and it will still be drawn. However moving it yet another file across turns it in to a win for White as the King can sneak around the other side of the pawns.
One other surprising example given was in the second diagram. Although it looks as though Black has no hope of holding back the King, Rook and Pawn, he just has enough to stop the pawn from queening. (The obvious 1.c7 is met by Kb7!)
You can probably discover other such positions using an online tablebase, but I'm not sure anyone has catalogued the full list of these patterns. I would be interested in hearing if anyone has.


HeinzK said...

It's not "injustice" because the quality of the white pieces is insufficient for a full point.

Anonymous said...

Most of the time it would *feel* like an injustice... after working so hard during the middle game for a 'winning' material advantage, given that these positions often with neither player having any knowledge of them. Peter