8. Objectives
Maximize the probability/expectation of a good outcome
Mimimize the probability/expectation of a bad outcome
Stop early because things are looking bad
Stop early because things are looking good (less common)
22. Dangers
Fitting complex models with little data
A priori overfitting
Exploring high dimensional design/outcome space:
Arbitrary simulation scenarios
23. Dangers
Fitting complex models with little data
A priori overfitting
Exploring high dimensional design/outcome space:
Arbitrary simulation scenarios
May have blind spot
24. Dangers
Fitting complex models with little data
A priori overfitting
Exploring high dimensional design/outcome space:
Arbitrary simulation scenarios
May have blind spot
Temptation to cherry pick scenarios