For AVERage=R-6, the distance between selected sets of atoms is averaged according to
where 
runs
through all possible combinations of distance restraints
between atoms ``
" in set 1 and atoms ``
" in set 2.
For AVERage=R-3, the distance between selected sets of atoms is averaged according to
where 
runs
through all possible combinations of distance restraints
between atoms ``
" in set 1 and atoms ``
" in set 2.
For AVERage=SUM, the distance between selected sets of atoms is computed by adding up single contributions
where 
is specified by the MONOmer statement.
The scaling by 
is required, in combination with the SUM averaging
option, to scale the distances corresponding to ambiguous
peaks in symmetric multimers.
The difference between the R-6 option and the SUM option
is subtle, and is best illustrated with an example:
if two distances are involved in the average, say 3 and 10 Å,
the R-6 average will produce an effective distance
of 3.37 Å, and the R-6 sum a distance of 2.99 Å, which is
probably the desired result for ambiguous NOESY crosspeaks.
For this reason, the SUM option
should be used rather than the R-6 option to treat ambiguities
for observed NOEs.
For AVERage=CENTer, the distance between selected sets of atoms is set to the difference between the geometric centers of the atoms,
