ABSTRACT
No single curve type provides the ideal framework about which all statistical problems can be formulated. For a continuous variate, at a specific point xi, it is curve evaluation F (x i), not f ( x i), that equals a probability. On the other hand, even though the cdf of a continuous variate is connected more directly with the important concept, probability, than its pdf counterpart, since no parametric model for any Normal cdf Φ exists, the “Gaussian” F cannot be computed with perfect accuracy. This is true even though models for all Normal /-type curves exist and can be expressed concisely enough for f (x i ) to be computed easily at any x\.
