Game Development Reference
In-Depth Information
We can lay out this data in a similar fashion as we did with our dentist data. Just
as we based the sizes of the “dentists' recommendation� buckets on the ratio of
those recommendations, we construct our buckets based on the relative sizes of the
four segments of the “guesser� population. When we lay them end-to-end, the total
width is 100. Because the figures were percentages of the whole, and we have ac-
counted for all of the groups that make up the whole, it makes sense that they add
up to 100. (We will find later that this is not a necessity.)
Once again, by laying the buckets side-by-side over our
x
-axis, we can deter-
mine the edges of the buckets (Figure 12.2). By dropping our metaphorical ball into
the buckets (by generating a random number between 1 and 100), we determine
which population segment our next guesser is going to represent. Theoretically,
63% of the guessers are going to be semi-logical, 30% will be random, and so on.
While the relative frequencies of the “33� and “22� guessers are small, there still is
a possibility that our ball will find its way into one of those two buckets.
FIGURE 12.2
The buckets created by arranging the relative population
segments of the Guess Two-Thirds Game. Note that the proportional sizes
of the buckets persist regardless of in what order we place them.
Notice that the order that we place the buckets in doesn't matter. In the bottom
half of Figure 12.2, we moved the buckets into a different arrangement. However,
because the sizes of the buckets haven't changed, the odds of our random ball drop-
ping into any one of them do not change either. For example, there is still a 4%
chance of a “33� guesser appearing.
