The smallest size would actually be a bit larger than 738/A320 - call it Keesje A320.5 in Capacity.
The most economical version (CASM) would actually be NSA-L - a little bit larger than than A321 for the majority of Short haul missions under 1,500 nm. It would have less range than A321 as it would be the smaller winged/lighter version. Which means lighter gear, lighter engines, smaller/lighter less draggy wing than A321.
The commonality would be so much with the larger wing version that it would no big deal to have both in a fleet and sub in depending on the route. Most of the spares excepting for the engine parts/gear/wheels would be the same.
Really think of a tight light 7W. Imagine you are sitting in an A320 - the NSA/NMA fuselage would be less than 18" wider to each side of you - hold your hands up 18" apart - it is not that big of a difference. The crown height would be similar.
If you build it to compete with the A321 then you build the wrong aircraft. A tight light 6W would kill a tight light 7W if both compete in the same segment (200pax).
There was a reason Boeing built the 757 and the 767. The 767 was a light 7W for its time but a 757 was still needed to cover the lower end, because a further shrink of the 767 was not viable. If the NSA-L is the most economic (as you say) then the short version will be around 10 percent less economic per seat if it shares the same gear and just a shorter fuselage (compare A319 to A320). In that case the old A321neo might be over all cheaper to use on said missions because it only costs 40mio or less to buy (in 2028 massive discounted to break the new short NSA 7W, which probably will be 60m-70m) and is only marginally more expensive to run.
Yeah the NSA-L and XL are CASM beasts then but if you need anything lower than 240pax single class, hello Airbus. Good you might be able to snatch up some cheap 737-8.
You are still thinking that an Ovalish 7W is radically larger than a 6W round fuselage. It's hard to say one would kill the other in terms of weight and drag per seat. The fuselages will not be significantly different.
Yes the NSA-S would be less efficient than NSA-L - so is the 738 vs 739 or 320 vs 321. NSA-S wins on trip costs.
A tight light 6W (737 size) won't work for A321 and above.
A321 size seems to be the winning size now - how big will it be 10 years from now when NSA could enter service?
With an optimized wing/gear/engines an 3,000 ish Range 7W NSA should be a lot more efficient than an 4,000NM+ A320/321, especially if they rewing the A320/A321 for even more capability - but that is what the big wing NSA/NMA is for.
Lets calculate a little just to see what efficiency difference will be:
B737 MAX 200has officially a capacity of 197 seats (28' pitch). That is the most efficient use of floor space possible (for the theoretical best economics).
Cabin length is 30m (33rows, 198 seats for easy calc), total length is 38m, width is 3.54m, total width is 3.76m
For a possible two aisle high density 197 seat model (the smallest possible version) you get:
Cabin length is 25.5m (28rows, 196 seats for easy calc), total length is 33.5m, fuselage height is 3.76m, with is 4.5m
Now we need to make assumtions: Two options new singe aisle and new double aisle have the exact same materials and technology so there is no advantage in this, it is only the different shape that counts.
Frontal area of a circular fuselage with 3.76m diameter: 11.1m^2
Frontal area of a oval fuselage with 3.76m height and 4.5m width: 13.29m^2
That is a 20% increase in frontal area, simplified of course but assuming the same shape for the nose it is a roughly 20% increase in area exposed to the air flow (angled). Now for skin drag:
Area of a cylinder with length 30m and diameter of 3.76m: 354.4m^2
Area of a eliptical cylinder with length 25.3m and 3.76m height and 4.5m width: 332.2m^2
That is a 6.25% reduction in surface area of the cylinder. Unfortunately as this are optimized designs the shorter aircraft will need a bigger tail structure to compensate for the lack of moment arm. So the surface area advantage will be even smaller.
So overall we will have a drag penalty for the twin aisle.
On top of that there will be a weight penalty from the bigger tail structure, and the need for reinforced structures due to the ovoid design.
Also due to weight disadvantage the twin aisle design needs either reinforced wings (if they are the same as on the single aisle) to carry the higher load or bigger wings to keep wing loading the same. This leads to higher weights.
All in all a super light efficient single aisle might have an empty weight of 40t, then the twin aisle will probably be in the range of 50t. So for the same mission profile the aircraft has to carry more structure with more drag. To compensate that you could have a better bigger wing but this advantages only counts on longer routes. So for missions <2500nm I can not see how this aircraft will ever be competitive (and that is the biggest segment today).
Now we can look at a stretch: 240 seats, same fuselages just longer:
Single aisle 40 rows, 36.5m cabin length
double aisle 34 rows (238 seats), 31m cabin length
Area of a cylinder with length 36.5m and diameter of 3.76m:431.2m^2
Area of a eliptical cylinder with length 31m and 3.76m height and 4.5m width: 403.83m^2
So the advantage of surface area is now 6.8% but the 20% increase in frontal area is still there. So it becomes better. Also for weight as the singe aisle will also need to be reinforced for strengthening the cabin and wings for the higher loads while the double aisle will be a bit better because the proportions are better (shorter), but it still has a too big tail structure inherited from the short version.
In the end the aircraft is still heavier probably not 10t anymore but around 5t, and has more drag even in this configuration. A single aisle has still the advantage. The only way to overcome that problem is to have more seats and more range.
There is physically no way you can build a twin aisle with the same seat count as a single aisle that is more economic. A 5 seat BMWX5 never beats a 5 seat BMW 5er in efficiency, you would need a 7 seat BMWX7 filled to do that on a per seat count.