On prediction and chaos in stochastic systems

We propose a new measure of sensitivity to initial conditions within a stochastic environment and explore its connection with nonlinear prediction and statistical estimation. We use modern statistical developments to construct and illustrate pointwise predictors and predictive intervals/distributions.