helios.voter prioritizes delegators first, and then votes randomly throughout the community when no burns are active. It's been a rare occurrence for burns to stop, ever since we set up the backup node switcher, mainly because we don't have any downtime related to hive-engine outages or rely on any single node as a point of failure. However, it's always a possibility to have time periods with no burns in the future. Good question, thanks for asking!