We found about a

40ms

improvement from

ranged_index

although we’re not completely finished with our testing of that yet