Why each test results are not the same?

I’ve been using webpagetestAPI for a while now,I wrote a small program using webpageAPI to detect the loading time of each page and other indicators,but sometimes the results of each test will be very different(Including ratings).So I think in the end how the measured value is accurate,I need an accurate set of values as the base value,can i do this? Or I still understand the error.
Need everyone’s help!!!