Important scenarios will be outlined in the tech design and may include:

• “Worst case” scenarios like large interval size, shallow index, varying array lengths, etc
• “Best case” scenarios like correlated predicates using a compound index, data with arrays of similar length, large (deep) index, small interval size, empty intervals
• Test for stability of estimates (how does accuracy change as writes occur)
• Different shapes of B-tree (closer/further from “balanced” or uniformly distributed)