Game Development Reference
In-Depth Information
The happy part is that, once established, these formulas can be applied to any
values we want to use. For example, simply changing the percentages of computer
failure will change the results significantly.
Figure 7.5 shows a slightly different scenario. In this case, the computer is a
complete piece of garbage and only has a 50% chance of surviving that first year. It
will need a repair 30% of the time (remember, those cost you $200 without the war-
ranty), and 20% of the time it will need to be replaced. By changing those figures,
we get the following:
In this case, we find that
E
(
W
) >
E
(¬
W
). Therefore, the expected value of
buying the warranty is greater than the expected value of not purchasing it. (Has
anyone yet considered not purchasing this computer at all?)
Another scenario, shown in Figure 7.6, takes into account the possibility that
the warranty is simply over-priced. Let's return to the original assumptions about
failure rates (20% = needs repair, 10% = boat anchor). However, we are going to
reduce the price of the warranty from $300 to $150. (Apparently there's a half-price
sale on warranties for junky computers.)
FIGURE 7.5
By increasing the percentage chances that the computer requires repair
or replacement, the expected utility of purchasing the warranty increases as well.
