A model in which the latest value in a time series is used as the forecast for all periods in the forecast horizon. Alternatively, it is a model stating that the difference between each observation and the previous observation is random. See naive model. Statisticians define the term as follows: A random walk is a time-series model in which the value of an observation in the current time period is equal to the value of the observation in the previous time period plus the value of an error term from a fixed probability distribution. It is a special case of a martingale.