[time-nuts] question about multi-way measurement

[Chris Howard]

I see the different forms of deviation measurements and they are all 
one-to-one comparisons.

Is there anything to be learned from doing mass data gathering?

For example, if I had a device of relatively good resolution that would 
let me
timestamp the events from 100 different clocks, then questions about the
change of the mean of the cloud of events, distance from the mean of 
individual
events, etc. could be obtained.

One of many things I have learned hanging around here is that some
very very smart people have already thought of anything that
might come to me.

It seems like, if there were a significant number of clocks involved, 
the mean
of the cloud of events would help cancel out positive and negatives and 
particularly
remove the short term randomness ?

So, has this sort of thing been done?
Why is everything one-to-one only?