Problem of fitting expected utility model in risky choice

Can you provide a runnable example? It will be easier to help if someone can run it and see the problem you are seeing.