Give one concise reason that any numeric estimate of the average value of a continuous positive function on an interval, obtained solely from a finite list of its sampled function values on that same interval, must be no less than the smallest listed value and no greater than the largest listed value.