and here you are 180 degrees out of phase with the real world of aerodynamics, canards are always making positive lift, (a good thing), and in the process making drag! (a bad thing), so the canards are inherently more "draggy" than the tail feathers. Now while the canards are making more lift, they are in "clean air" and are likely more effective per size and moment arm than tail feathers which reside in "dirty air"?
you correctly assert that the main wing is able to "mask" the tailplanes, while the canards are standing out front.... (they may "mask" a portion of the main wing head on, but at more oblique angles, not so much), they will likely raise RCS from the frontal and possibly side angle.
the J-20 does not have "all moving" tailfins, the F-22 does have all flying stabilators, which are also used to roll the aircraft
in the end, its six of one, half a dozen of the other, both arrangements have strengths and weaknesses.
Tails generate positive lift and drag too. You don’t really think air decides to skip flowing over the tails because they’re tails do you? Yes, that positive lift might be smaller per area because of downwash from the wings, but that’s why tails tend to be bigger, to generate the same amount of pushing force needed to pitch the plane. Meanwhile that drag from being an object that interacts with the air stream is still there. In fact with that combination of being just as present to the air stream and typically larger size due to control demands it could be argued that tails in general generate as much drag as or more drag than canards (to be clear I’m not saying that a plane with tails is always draggier than a plane with canards, as always I dislike drawing hard conclusions about specific designs based on superficial cherry picking of individual features).
I’ve also given this canard stealth thing some thought, and I’m not sure the prevailing conclusions are correct. They sound more like erroneous extrapolations of some laymen heuristics. If we think about how RF backscatter against a wing works, the incidence angle the wave encounters is what determines their angle of reflection. Against a plane’s surface this is determined by sweep angle, which is true for both the canard and wing. If their sweep angles match they should be reflecting the RF wave in the same direction. In this sense the general amount of RF energy reflected back to the source for a canard surface is at most a linear addition to the amount reflected back by the wing when the canard edge doesn’t block direct line of sight to the wing, and a non contributor in the general sense when the canard edge is coplanar to the wing edge, at least in terms of total energy. This may mean that in terms of total possible energy a canard+wing is reflecting the same amount of total energy as a larger wing, as more surface always equals more energy reflected, but when that is the case the extra contribution of surface from a canard, like with a larger main wing, shouldn’t be noticeably significant because backscatter depends on *area*, and edges for small surfaces don’t have a lot of that, and in addition that area is *effective* area, which isn’t defined by physical dimensions alone but by how reflective the physical object is to the RF wave, which should be much less for surfaces treated to dampen RF reflection.
Despite all that, the typical argument against using canards comes in three flavors. The first is that as an additional object exposed to the RF wave canards will generate their own dipole lobe (the beam of RF energy reflected away from the source) separate from the wing. However, so long as the canard shares the same sweep as the wing the angle of their dipole lobes will be the same, which is to say they should be reflecting RF energy in the same direction, which then therefore means the canard’s extra dipole lobes shouldn’t disrupt the “bowtie” shape generated by the wing to forward facing RF waves. A canard with the same sweep as the wing should therefore at worst only have negligible extra contributions to frontal RCS, and at best make no difference (with the dipole lobe of the canard being totally indistinguishable from that of the wing and the rest of the fuselage), except when extra exposed edges contribute some extra area for the RF wave to interact with, which itself should be minuscule for treated surfaces.
The second kind of argument is the traveling wave problem, where as a function of being in front of the main wing the traveling wave that propagates along the canard’s surface exits from the canard’s trailing edge and interacts with the main wing in a more scattered fashion, which makes that portion of RF reflection more difficult to control. If the canard is either reflecting or dampening most of the RF energy hitting it though the additional contribution of RF energy propagated to the portion of the wings behind the canard should be quite small, smaller still if that portion of the wing is treated to dampen RF reflections, and even smaller if the canard also has their trailing edge treated to dampen RF reflections. Furthermore, the angle at which those traveling waves are reflected once they exit the trailing edge may be effectively controlled by metamaterials to reflect that energy in a more favorable direction and further limit their interactions with the main wing. In the J-20’s case, the canards being angled at a dihedral rather than being coplanar may actually help reduce this component of RF reflection. The trade off is of course more of the wing edge behind the canard is exposed, so you get that linear addition I mentioned earlier, but again that may end up being negligible.
The last kind of argument is that the canards, by adding an extra wing root, adds an additional corner reflector. In the J-20’s case though the canard wing root is coplanar with the main wing, so there should only be one corner reflector facing forward. Of course all these arguments apply strictly to forward facing RF waves. For azimuths that look more sideways or vertically, canards may actually be a bit better than horizontal surfaces, since at those angles line of sight to the tails aren’t being blocked by the main wing, and tails tend to be much larger than canards.