Game Development Reference
move 30. The fox has just forced the hounds to break their line. Move 30 might
seem strange, but recall that when the hounds first break their line, they do not
look ahead. Instead, they pick the move that puts the fox farthest from the hole.
The alternative of moving the left-most hound would have created a shorter path
to freedom for the fox.
While move 30 might appear to be strange, move 32 at first glance appears
completely insane. Before they made move 32, the hounds looked ahead,
expecting the fox to always take the shortest path to freedom. They saw that they
could put their line back together at move 34, yielding a board with an evaluation
score of 116. They also saw that they could put their line together on move
36 with a score of 117. What they liked best was putting their line together on
move 42 with a score of 119. With this evaluation function, the hounds appear to
have mastered the concept of delayed gratification.
If the fox moves as the hounds predict on move 33, heading downward for
freedom, then the hounds will toy with it. They will open a second hole to tempt
the fox, as shown with move 34 in Figure 6.8. If the fox takes this new bait, the
hounds close the new hole on move 36, setting up the inevitable enclosure shown
a few moves later in move 42. If the fox ignores the new hole opened on move
34 and heads along its original path downward toward freedom, the hounds will
be able to block both holes before it escapes (not shown).
The critical move for the fox was move 33, visible in board 34. Is there a better
alternative for the fox after the hounds make move 32, as shown in Figure 6.7?
The hounds toy with the fox before crushing it.