Problems as Must Win Situations

There are weaknesses with solving tactical problems. The most obvious, is that when you are presented with a problem position, you know there is a solution.

The problem of the day from chessgames.com (should be in a widget on the right) sometimes does not have a "solution", and if there is a solution, it is given by the commentors, some of whom are very good. (Once's posts are always well worth the time spent on the problem.

Another problem is that you often know whether you are playing to win or draw. The problems at Shredder's page do not have this issue. They are just find the best move, and sometimes they are wins and sometimes they are draws.

Problem 43918 from Chess Tempo has this second problem.

Black to play and win

It is not too hard to find the draw starting with 1...Nxf2, but seeing the win is much harder. But seeing the only move that can win, knowing that there is a win? Not so hard. 1...Nxc3 is the only way to win. It might lose, except that this is a problem and you know there is a win. After 2.f6, again there is only one possibility for a winning move. All moves but 2...Ne4 are sure losers, so that must be the move. After 3.f7, the only possible winning move is Nxf2, as there is no checkmate, and no hope against Q+R with only N+B and pawn. After f8/Q, setting up the d-pawn to queen with Ne4 is again the only possible winning move, for which the site awarded me the solution, and a nice +3.5 rating.

In a game, I would only have played 1...Nxc3 in a must win situation. I had good discipline on this problem, and though I knew that 1...Nxc3 was the only possible winning move,  I still spent 30 minutes on this problem, working through the variations to find the path to this position:

White to play

Black will need to find shelter for his King from the checks, and keep some of his pawns. Then White will have to take the Ne4 with his Q leading to ...d2+ and ...Bxe4, and a won endgame for Black.