When asked how many moves ahead Kasparov can think, he replied that it depended on the positions of the pieces. "Normally, I would calculate three to five moves," he said. "You don't need more.... But I can go much deeper if it is required." For example, in a position involving forced moves, it's possible to look ahead as many as 12 or 14 moves, he noted.
The normal GM searches a few moves per second, while a chess engine searches millions per second. See the difference?
Search()->
30+ conditions for root.
Conditions:
What type of move are you?
1) Developing (including castling)
2) Exchange
3) Sacrafice
4) Fork
5) Pin
6) Discovery (check, attack)
7) Skewer
8) Tempo
9) Mate (Threat)
10) Draw (Repeat)
11) Mobility
12) Positional
13) Attacking
14) Defensive/Protective
15) Capture
16) Open rank
17) Open file
18) Open diagonal
19) Castled or Not Castled or Capable of Castling
20) Castling stopping
21) Controling squares
22) Attacking squares
23) Free squares
24) Occupied squares
25) King Safety
26) Safe/Retreat squares
27) Supporting
28) Attack Supporting
29) Capture Supporting
30) Pin Supporting
31) Center Control
32) Other
Order your root move based on the total count of conditions met, bonus, penalty, etc.
To be more successful at Chess, an engine or individual should take these conditions into consideration for a move. Planning is a key to winning. Design & develop a search for just the 3 basic attacks would be a good place to start (forks, pins, & skewers). This will lead to the removal of piece table scores, and if done correctly, alpha/beta. Also, more human-like play will occur.
This would be a step forward for chess.
