Stormcloud,
The simple answer is that the CG position in relation to the datum is irrelevant other than in regard to the actual calculation itself. The following may be useful in explaining this statement.
Some background information, applicable to any aircraft -
(a) a manufacturer will publish station data based on convenient datum locations chosen for the convenience of the design and manufacturing organisations. The convenience of the operator is of little or no concern at this stage.
(b) there is no intrinsic value in any particular datum locations and they can be chosen on a basis no more rigorous than the whim and/or preference of the design organisation, except that lateral arms normally are measured from the aircraft centreline. In effect, the various datum locations can be viewed as being no more than nails hammered into the aircraft from which one can hang a tape measure to measure distances to any loading location of interest .. no more, no less.
(c) so far as the longitudinal arms are concerned (and this is what you are considering), the datum can be located anywhere along the aircraft from somewhere out in front of the nose through to somewhere out behind the tail, although I can't think of any aircraft where the manufacturer has chosen a datum aft of the wing.
(d) when it comes to doing pilot type loading calculations, there are two principal considerations
(i) if the CG calculation is to be done by doing a bunch of sums (either longhand or within a computerised reservation and loading system) it doesn't really matter what the datum location is (albeit that, for longhand calculations, the choice of datum can materially affect the ease of manual longhand calculations)
(ii) if the calculation is to be done manually using a graphical loading system (usually a trimsheet of some sort or other) then the errors implicit in the design and operation of the system will be minimised if the datum chosen FOR THE SHEET (and this datum can be quite different to that used by the manufacturer for design purposes) is located somewhere near or inside the CG envelope.
For light aircraft, the best location typically is the aft limit and for heavy aircraft the preferred location will depend on the shape of the envelope. Typically, in this latter case, it will end up somewhere inside the envelope.
For larger aircraft, the manufacturer often nominates a sensible second datum for trimsheet use, typically called the trim datum or some similar term. There is no particular need for the operator to use the manufacturer's nominated datum and, in many cases, the operator's trimsheet will be based on yet another datum chosen for the operator's particular reasons. Apart from a consideration of errors in the design and execution of a trimsheet, there is no reason why one datum position is better than another.
. .So it follows that the CG location for flight purposes has no significance with respect to the datum other than to the extent that the calculated position has to be within the relevant limits specified either in the AFM or, more usually, the constricted limits published in the trimsheet.
I suspect what your question is trying to determine relates more to the desirable CG for flight. If this is the case, then the main loading emphasis ought to be directed toward achieving a CG which gives minimum trim drag for a significant portion of the cruise.
The manufacturer usually provides some guidance in the loading documentation or, at the very least, in information provided for the use of the operations engineering people. Typically this turns out to be somewhere toward the rear CG envelope limit.
For longer range aircraft, attention to trim drag in load planning can add up to quite a significant cost variable. For instance, I am aware that one 744 operator which used the aircraft on payload limited sectors looked at a revised fuel use schedule to see if an improved CG control in flight and its effect on trim drag could be converted into sensible dollars. The end result was something like a sector fuel saving in the region of 150 kg (ie an extra paying customer or two) if my recollection is correct. On an annualised basis this represented a potential profit increase in the order of USD 1.0M ... not to be sneezed at, I would have thought. Even if my recollection of the specific numbers is a little flawed, the principle remains valid.