Random walk is used in the analysis of time series.
If a time series is a random walk, then future values cannot be predicted based on past values, even if there seems a correlation between subsequent
points in the time series. A random walk is different from random points, because every new point is based on the previous and then a random change.
In time series prediction it is important to make sure that we don't have a random walk.
Very detailed explanation in this MIT course:
