Game Development Reference
In-Depth Information
the original assumption we made about our partner… that we could have
no idea
what it was that he would do. That may or may not be the case—it depends on our
knowledge of our partner. It could still very well be wise to betray him if we have
reason to believe he is going to do the same to us. If we believe he is not only intel-
ligent enough to recognize the hidden “out,� but also thinks enough of
us
to assume
that
we
would take it as well, then we can choose to keep quiet and meet him for a
cheeseburger in six months. We will examine this more in Chapter 6.
I
NTHE
G
AME
Dueling Rocket Launchers
To see how this sort of game theory staple is important, let's put the Prisoner's
Dilemma into a hypothetical game setting. To design a scenario that accurately re-
flects the quandary of the Prisoner's Dilemma, we need to look at the parameters
that surround the choice. First, we must have two parties with two identical choices
facing them. Second, we must have a conservative choice and an aggressive choice.
Both of those choices must have a reasonably positive outcome and a somewhat
negative outcome based on how they match up with the other player's choice. That
is, we need to have a 2
2 scoring matrix where the positives and negatives roughly
match those in the original Prisoner's Dilemma.
Imagine we are in a shooter type environment as shown in Figure 5.3. Both
players start in a position of being mostly hidden from the other player. The cover
we are behind protects us from anything
except
rockets. If a rocket were to strike
our hiding place, we would be killed. We are fully healthy but lightly armed. We can
fire on our opponent from our cover, but only do very light damage to him.
Likewise, if he stays hidden, he can do light damage to us as well.
×
FIGURE 5.3
Either agent could elect to run into the open and grab a rocket launcher
or to hide and wait for reinforcements.
