02.03.2013 Views

Downloadable - About University

Downloadable - About University

Downloadable - About University

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

224 Revising judgments in the light of new information<br />

his belief, or where the prior probability is very small, implying that he<br />

strongly believes that gas will not be found. In the latter case, so strong<br />

is his disbelief that he severely restricts the effect of the disconfirming<br />

evidence from the test drilling.<br />

At the extreme, if your prior probability of an event occurring is<br />

zero then the posterior probability will also be zero. Whatever new<br />

information you receive, no matter how reliable it is, you will still<br />

refuse to accept that the event is possible. In general, assigning prior<br />

probabilities of zero or one is unwise. You may think that it is impossible<br />

that San Marino will become a nuclear power by the year 2020. However,<br />

if you hear that seismic instruments have detected signs of nuclear testing<br />

there then you should allow yourself some chance of being persuaded<br />

by this information that the event might just be possible. Assigning a<br />

very small but non-zero prior probability might therefore be wiser.<br />

Ironically, if the new information has less than a 0.5 chance of being<br />

reliable its result is of interest and the more unreliable it is, the greater<br />

the effect it will have on the prior probability. For example, if the test<br />

drilling is certain to give the wrong indication then you can be sure that<br />

the opposite of what has been indicated is the true situation!<br />

Applying Bayes’ theorem to a decision problem<br />

We will now consider the application of Bayes’ theorem to a decision<br />

problem: a process which is sometimes referred to as posterior analysis.<br />

This simply involves the use of the posterior probabilities, rather than<br />

the prior probabilities, in the decision model.<br />

A retailer has to decide whether to hold a large or a small stock of a<br />

product for the coming summer season. A payoff table for the courses<br />

of action and outcomes is shown below:<br />

(Profits)<br />

Decision Low sales High sales<br />

Hold small stocks $80 000 $140 000<br />

Hold large stocks $20 000 $220 000<br />

The following table shows the retailer’s utilities for the above sums of<br />

money (it can be assumed that money is the only attribute which he is<br />

concerned about):

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!