When calculating the variance of a project's completion time, why is it that the 'most likely time' is not factored into the equation, especially since it carries the most weight in calculating the expected project completion time? I understand that it measures the dispersion of the activity time, but wouldn't there be an effect from the 'most likely time'?