The interference problem for the betting interpretation of degrees of belief