Select as many as apply — they shape the scoring.
Outputs are estimated scenarios, not guaranteed predictions.