A cost measures how undesirable an outcome is. It can be treated as the negative of utility: higher costs correspond to worse outcomes. In decisions involving interventions, subtract implementation cost from the utility of the result to find net utility. A machine-learning loss function is related, but typically measures mismatch between predictions and observed data. section 12.1
Choosing a policy
A job-training policy with cost 1, expected utility 7.5, and net utility 6.5 compares favorably with a riskier policy costing 2, whose expected utility is 3 and net utility is 1. The second policy’s larger possible payoff does not make it the better choice. section 12.7 section 12.7
Table 12.3 Utility of two available policies under possible scenarios and the corresponding probabilities
|
Option
|
Policy A
|
Policy B
|
Doing nothing
|
||
|
Cost
|
1
|
2
|
0
|
||
|
Utility
|
10
|
5
|
30
|
0
|
0
|
|
Probability
|
0.5
|
0.5
|
0.1
|
0.9
|
1
|
|
Option
|
Cost
|
Utility
|
Probability
|
|
Policy A
|
1
|
10
|
0.5
|
|
5
|
0.5
|
||
|
Policy B
|
2
|
30
|
0.1
|
|
0
|
0.9
|
||
|
Doing nothing
|
0
|
0
|
1
|
Compare the policies’ costs, possible utilities, and outcome probabilities. section 12.7
















