Lately all of my solutions have been in very low percentiles with no notable differences from solutions in the higher percentiles.
There should be a better way to isolate the runtime and memory measurements.
Have others been noticing this trend?