These tasks combine methods you have already used on new fictional records. The output is a decision record as well as a notebook result: it should make the evaluation population, observation clock and error units clear. Start by identifying which rows belong in the fit for the stated prediction time. Then choose the calculation that matches the question. No new estimator is required. Several tasks include a boundary, a missing outcome or a limited queue because those details change the meaning of an otherwise familiar score. Preserve the requested output names so the checks can inspect your reasoning through the objects you produced.