Rank the backlog
Ordering uses lexorank strings, which means moving one item rewrites one row rather than renumbering the list. On a long backlog that is the difference between a reorder that is instant and one that is not.
Commit, then start
Pull issues into the sprint, set the goal, and start it. The commitment is snapshotted at that moment.
Estimation uses a scale you configure per project, so a team that points in Fibonacci and a team that estimates in days are both first-class rather than one of them working around the other.
Read the charts honestly
Burndown shows the sprint against its own commitment. Velocity is computed from completed sprints, so it is only as meaningful as the number of them you have finished.
Cumulative flow comes from daily rolled-up metrics rather than from a live query, which is what lets it show a month without scanning every issue every time you open it.