We have built models, filtered states, and estimated parameters. Now the real test: can our model say something useful about the future? This chapter shows how to turn a fitted linear state‑space model (a model with hidden states that evolve and produce noisy measurements) into sharp forecasts, realistic uncertainty bands, and honest model checks. It is the moment all that earlier work pays off. It is simpler than you might think once you see the two main paths: quick analytical updates with the Kalman filter (a recursive algorithm that estimates the hidden state from noisy data), and full probability forecasts through Monte Carlo sampling (using random draws to approximate distributions).