Sometimes tactical combinations are short and the outcome is fully foreseeable - for instance if it is a 2-3 move deep combo. Other times the outcome of the combo is far deeper, as in the following game. Here Surbiton player Angus James is White against a Norwegian player in the London Chess Classic FIDE Open 2013.
Black is about to break through on the queenside and wipe out White's pawns. After a lot of calculating I find the only way to effectively breakthrough on the kingside, and take the plunge with a piece sacrifice.
What was particularly pleasing about this game is that it turned out the piece sacrifice 19.Nxf7 worked against any defence - although I didn't know that when I played it. I tried hard to make it work in all lines in my calculations, and in the end I stopped calculating and just played the move. I knew it was the move that was most natural in the position - sometimes that is enough. We have to make decisions based on incomplete information, and we only discover if we were right as the game is played out over the board.