The set of numbers you give is real. A set of estimates are not real, they are estimates. You can not add up a bunch of estimates, divide them by 10 and then pretend you have any sort of exact number, certainly not more exact than the original estimates.
In science you work from exact to exact, not estimate to exact.
Just use your head, how can you rank a set numbers based on rounding off the original non data to a set of estimates, averaging them to produce a mean and then say that 3.05 is higher or better than 3.01? Sorry, you can call me all the names you want, but inexact numbers can only produce inexact averages and/or rankings. You are using the so called data to produce WOW differences which the 'data' can not support.
In science you work from exact to exact, not estimate to exact.
Just use your head, how can you rank a set numbers based on rounding off the original non data to a set of estimates, averaging them to produce a mean and then say that 3.05 is higher or better than 3.01? Sorry, you can call me all the names you want, but inexact numbers can only produce inexact averages and/or rankings. You are using the so called data to produce WOW differences which the 'data' can not support.