When I read endgame texts (bleah...) by Dvoretsky/Yusupov, frequently, the authors describe schematic planning in stages outlining the general ideas leading to the final win/draw and how the ideas pan out in the actual games. In a recent fbchess game played between FM Ong Chong Ghee and IM Terry Toh, during the post mortem, I suggested a possible drawing idea and the latter indicated that he had anticipated the possibility and had set out an 8 stage plan(!) to overcome it.
This was the position we analysed during the post mortem. My drawing idea was to put the king on a2 and when the black king attacks the Bb7, put it on a8 and wait for Black to play ...c5 when the bishop attaches itself on d5. Even the loss of the d3-pawn might not matter. For example, if White manages to reach the following ending,
there is still no win, since the bishop is capable of defending any pawn attacked.
However, back to the original position analysed,
Terry outlined the following plan to win:
1) Play the king to b6 to prevent White from playing a possible c4-c5 pawn sacrifice.
2) Put the h-pawn on h6 to prevent h5-h6 and the possibility of the h7-pawn being swiped by the light squared bishop.
3) Play ...c6-c5 to secure the bishop on b4.
4) After taking the b2 and a3 pawns, push a4-a3 to secure the pawn with the Bb4.
5) Round up the d3 pawn with the king by zugging the white king with the aid of the Bb4.
6) Walk the king to g5 to hit the h5 pawn, forcing White to play Bf7 to hold it.
7) Play the ...d6-d5 pawn sacrifice, forcing White to allow a passed c-pawn.
8) Get the king over to the queenside to aid the c-pawn home.
Let's see how it pans out.
Now that's what I call serious schematic thinking in the ending!