Game Development Reference
In-Depth Information
FIGURE 2.4
There is a bias toward using the first and
last blades in a pack longer than the other three.
Just as my relatively lean college days led to a more significant reliance on the
final blade, other people could have other patterns. Someone who is better off may
not be concerned about moving on to the next refill tray, thereby not sticking with
that fifth one long enough to risk a medical emergency. Someone who has a specific
weekly pattern of usage may find he is
always
on the third blade before his “big date
night� and will switch to the fourth to get that “clean, close shave� the commercials
try to tell us is a strict requirement for actually being male. (I admit that I'm reach-
ing a little bit on examples now.)
Now, razor blade usage patterns may not seem terribly relevant to games, but
the illustration is more of the thought process involved than it is of the specific
example. There are differences in how a given individual approaches each blade, and
there are differences in how multiple individuals approach the whole blade-usage
predicament. The point is that many typical behaviors can be quantified this way.
What's more, the quantification can be done in such a way as to emulate the depth
and personality of the subject.
Turning Observations into Numbers
Observing the distribution of possible choices is only the first step, of course. AI
agents are there to act, not analyze. However, from a conceptual basis, the transi-
tion is not difficult. Using the razor blade example, if we proceed with the premise
that the figures in Figure 2.4 are accurate usage percentages, it would be reasonable
to assume that those same percentages would apply to the likelihood of me using
any one of those five blades at a particular time.
So, if I were faced with an entire tray of five blades that I had used before, and
the order has been preserved, you could make an educated guess as to which blades
I would be likely to select.
In this case, we know that blade 5 is likely to be used the most, followed by blade
4. Blades 2 through 4 are the least used of the batch and would therefore be the
most attractive to me. You could make an assumption based on these facts that
