Before the digital age when computing power was low with high costs, data were few and must be filled with human knowledge to build alternative low, base, and high case deterministic models.

The modeling processes were long and segmented. They led to numerical models and would forget about initial data.

​

In the digital age when computing power is high with low costs, Big Data sets are scanned by machine learning algorithms to bulld mathematical probabilistic models that instantaneously deliver P10 P50 P90  scenarios.