Chris-
I don't fully understand the question, I guess, but I'll try to contribute. Is your illustration from an R/C Proving Grounds calculator? I think I've seen it before, but I don't have a direct link to it. If so, there are a couple things to look out for. First, their static margin expectations are more attuned to R/C and full-sized aircraft, whose c.g.'s are significantly behind those used in CL. Also, the N.P. locations are based on a simpified analysis that I did not fully "get" and which also actually depends on vertical aft-wing placement and a lot more. I have pretty well given up on reasonable analyses, but there are some out there, including good papers from Stanford aeronautical. Anyway, their N.P. is not necessarily accurate, and for sure, you don't want to locate it as though the aft wing is as efficient as the forward one.
Here's what I think I know about Canard's in general (and agree with Ted for stunt). Not only is the "main" wing less efficient, due to down wash from the canard, but for reasons of longitudinal stability and stall behaviour, the canard must be more heavily loaded. The main wing ends up unable to use its entire lifting capacity. So the canard loses its intuitively apparent advantages and is less efficient in overall function than an aft tail. If computer aided control/stability is employed, then it becomes about as efficient as an aft-tailed plane.
From this cursory "analysis", I'd conclude that you'll want the c.g. further ahead of the plane's NP than shown, but that the NP may be a bit off anyway. These on-line calculators are great for wings, but probably not too sophisticated regarding aircraft NP's. I think they're great for normal stunt configurations, coupled with our normal rules of thumb, but we don't really have that luxury with canards. We already have the wing and tail a.c. approximations to deal with - the assumption that all points on a wing lift equally. We adjust for this, but it's tougher when you couple them together and then add a smaller lifting surface that compromises the larger.
So you may be correct about their proximity in this mathematical model, but the facts that it is an approximation applied to approximations, and that our c.g.'s are placed forward of their R/C predictions, makes the analysis harder. In reality, the NP may indeed be further forward than predicted (especially when predicted by assuming equal lifting ability per unit area, front and rear), but so is the CL c.g. and shifted c.g. needed to load the canard surface more heavily. Some reading and further researching are probably in order!
SK