In the “Performance Results (Median Run - SpeedIndex)” table, it selects a run for the “First View” and it selects another run for the “Second View”. What metric is it using to pick the media run?
Is it time to “first btye”, “largest contentful paint”, or “document complete”, or “fully loaded”?
It uses “Speed Index” (or whatever metric is next to “Median Run” in case you are running a private instance that uses a different metric.
What should be the ideal medianMetric when comparing between performance test results ?
@pmeenan While working on performance improvement the general workflow I follow is take before and after measurement. And one challenge I often face is say I have made fixes around improving LCP (added image preloading) but while comparing between two test results each having 5 runs say there is some difference in TTFB between default compared run (due to caching or any other reason). So in this case what should be correct medianMetric I should use ? LCP, TTFB or default speedIndex to get the correct comparison report ?