Answer: The exposed and nonexposed groups under study be as similar as possible with regard to possible confounding factors.
In cohort study you will divide the groups into 2. The group doesn't need to have equal risk, but you should make sure the group is as similar as possible.
Similar group has similar confounding factor.