For one thing, because you really can't (wavelength considerations alone
make that a ridiculous claim), and even if you could, you'd get a screen
full of noise.

Thats true,but only if you use radar as amplitute based classical binary
detection method (conventional radar).
But if you start to consider radar as a data carrier and process,for example
polarimetric data,you can easily filter out the noise.
Regarding ridiculousness of claim,this claim was the talk of the town last year
somewhere in East Coast.if* you know what you're looking for, and under
perfect conditions.

The only demonstrated multistatics have been working on targets
literally a thousand times the size of stealth planes (10 m^2 versus
0.01 m^2).

Lets say only sofar,the frontal RCS of both B2 and F22 is 0,0001 sqm but both
of them could be easily detected,tracked (and imaged) by US multistatic
system,which is optimized for the next generation stealthy cruise missiles and
UCAVs.
Why ? Because current RCS definition has a meaning only if you face
backscatterers,if you face multistatics thats different story.

While stealth designers also *absorb* energy, and use those much-reduced
reflections to make ECM much more potent. that's the big weakness of
multistatics, you know... *way* easier to spoof.

As I posted before,for meaningful backscatterer radar detection range
reductions you need an echo reduction at least in order of 10000 whereas the
best RAMs today offer only reduction in order of 100.
RAMs are a stealth designers Band-Aids,they use them only if everything else
fails,the weapon of every passive stealth platform designer is the "Hardbody
Shaping".
Probably you wont see any RAMs,RASs etc in the platforms of future because of
the development of HPM weapons.
If you face HPM weapons the last thing that you may want is to absorb their
energy.
Spoofing? Well it depends,if you know the location receiver lt might be easier.
But do Stealth planes need ECM support like lowly counterparts?