|visits||member for||2 years, 3 months|
|seen||Aug 21 '12 at 19:10|
Proxy for risk in portfolio theory when return can take only two values
OK. I get this. And I guess a similar, although less extreme, problem exists with a typical stock market application, as returns there aren't actually normal. I think I can adjust my formulation so that assuming the returns are normal is less of a stretch. How do you judge when your returns are too far from normal for the model to be useful, I wonder.