Overtime the variation that gives an average

result closer to the true value will tend to dominate, as repeated experiments will eventually converge towards the most accurate measurement. This is known as the law of large numbers, which states that the average of a large number of independent, identically distributed random variables will converge towards the expected value or true underlying parameter of the distribution. Thus, over time, the variation that gives an average closer to the true value will be more reliable and trustworthy.