Negamax is a common way of implementing Minimax and derived algorithms.

• Instead of using two separate subroutines for the Min player and the Max player, it passes on the negated score due to following mathematical relation:

NegaMax is called from a root function.

In the loop which generates all the root moves:

• Generate a Score relative to the side being evaluated, for a specific move.
• In the Root retain the returned score for each potential move.
• Pick the move with the lowest score.

max(a, b) == -min(-a, -b)

This is how the pseudo-code of the recursive algorithm looks like. For clarity move making and unmaking is omitted.

int negaMax( int depth ) {
    if ( depth == 0 ) return evaluate();
    int max = -oo;
    for ( all moves)  {
        score = -negaMax( depth - 1 );
        if( score > max )
            max = score;
    }
    return max;
}

http://web.archive.org/web/20120209042116/http://chessprogramming.wikispaces.com/Negamax