Thats what mem (he claims to be the inventor the collar) was doing before he disappeared from the forum. He developed a collar v2.0 :D.
I will add some links to this post…
Here men talks about the RGB-ratio of the collars mirror coating and the tint bin of the LED having an effect on the final performance-gain of the collar.
More talk: Wavien ceased operations - #47 by MEM
Here it gets interesting. He talks about boosting the XP-G2 to a luminance that would equate to 2900 lumens if it were this intense without the collar. The actual maximum of the XP-G2 S4 2B (the good old one) is 6A and 1120lm (see koef3s test ). This means that mems “v2.0” collar achieves up to ~160% gain. Thats about 30% more than the Wavien collar (120% gain). I find that believeable.
Here he mentions that collars made out of stainless steel don’t work very well because this material only reflects around 30% of blue wavelengths.