Differences

This shows you the differences between two versions of the page.

--- chess:programming:evaluation [2021/10/30 22:46] – created peter
+++ chess:programming:evaluation [2022/01/06 20:07] (current) – peter
@@ Line 12: / Line 12: @@
 ----
-[[Chess:Programming:Evaluation|Evaluation]]
+[[Chess:Programming:Evaluation:Tests|Tests]]
 ----
+===== Finding the Best Move =====
+Evaluate all the available moves using minimax() or alpha-beta pruning() or similar.
+Then returns the best move the maximizer can make.
+<code cpp>
+function findBestMove(board):
+  bestMove = NULL
+  for each move in board :
+    if current move is better than bestMove
+      bestMove = current move
+  return bestMove
+</code>
+----
+===== Checking for GameOver state =====
+To check whether the game is over and to make sure there are no moves left.
+<code cpp>
+function isMovesLeft(board):
+  for each cell in board:
+    if current cell is empty:
+      return true
+  return false
+</code>
+----
+===== Considering Depth =====
+Occasionally an evaluation function will return the identical value for two possible moves.
+  * But one of these moves may result in a slower victory or a faster loss:
+    * Move A could win in 2 moves.
+    * Move B could win in 4 moves.
+Even though both moves evaluate to the same score, it is usually better to select the quicker move, i.e. in this case Move A.
+  * To overcome this problem the depth value is subtracted from the evaluated score.
+  * This means that in case of a victory it will choose a the victory which takes less number of moves.
+  * In the case of a loss it will try to prolong the game and play as many moves as possible.
+So the new evaluated value will be:
+  * Move A will have a value of +10 – 2 = 8
+  * Move B will have a value of +10 – 4 = 6
+Since move A has a higher score compared to move B means select move A over move B.
+  * The same thing must be applied to the minimizer.
+  * Instead of subtracting the depth value, it is added as the minimizer always tries to get as negative a value as possible.
+  * We can subtract the depth either inside the evaluation function or outside it.
+<code>
+if maximizer has won:
+  return WIN_SCORE – depth
+else
+if minimizer has won:
+  return LOOSE_SCORE + depth
+</code>
+----
+Consider both hits and misses:
+  * evalHashHits
+  * evalHashMiss
+  * pawnHashHits
+  * pawnHashMiss
+----
+Eval is adjusted towards 0 if a draw is likely.
+Pawns are valued slightly less but becomes a bit more valuable in endings.
+Rook & Queen on 7th rank only awarded if enemy king on 8th or enemy pawns still on 7th.