Simplify KBNK endgame implementation

We do not need to change the winnerKSq variable, so we can simplify
a little bit the logic of the code by changing only the loserKSq
variable when it is necessary. Also consolidate and clarify comments.

See the pull request thread for a proof that the code is correct:
#1854

No functional change

src/endgame.cpp

@@ -120,7 +120,7 @@ Value Endgame<KXK>::operator()(const Position& pos) const {
 
 
 /// Mate with KBN vs K. This is similar to KX vs K, but we have to drive the
-/// defending king towards a corner square of the right color.
+/// defending king towards a corner square that our bishop attacks.
 template<>
 Value Endgame<KBNK>::operator()(const Position& pos) const {
 
@@ -131,18 +131,12 @@ Value Endgame<KBNK>::operator()(const Position& pos) const {
   Square loserKSq = pos.square<KING>(weakSide);
   Square bishopSq = pos.square<BISHOP>(strongSide);
 
-  // kbnk_mate_table() tries to drive toward corners A1 or H8. If we have a
-  // bishop that cannot reach the above squares, we flip the kings in order
-  // to drive the enemy toward corners A8 or H1.
-  if (opposite_colors(bishopSq, SQ_A1))
-  {
-      winnerKSq = ~winnerKSq;
-      loserKSq  = ~loserKSq;
-  }
+  // If our Bishop does not attack A1/H8, we flip the enemy king square 
+  // to drive to opposite corners (A8/H1).
 
   Value result =  VALUE_KNOWN_WIN
                 + PushClose[distance(winnerKSq, loserKSq)]
-                + PushToCorners[loserKSq];
+                + PushToCorners[opposite_colors(bishopSq, SQ_A1) ? ~loserKSq : loserKSq];
 
   return strongSide == pos.side_to_move() ? result : -result;
 }