Pretty much all implementations of RT in games so far have been completely connected to the frame buffer's resolution. Most effects are spawning rays for every render-target pixel, or every other one, or 1/4 etc. That means all those games RT tracing performance cost will be reduced "automatically" proportionally with resolution reductions*, in direct oposition to your assumptions that just because rays are traced in world-space they bare no correlation to render-res. Your point was simply wrong. Not just wrong, you were claiming EXACTLY the OPOSITE of what is actually a fact. And snarkly so.
*the exception being acceleration structure build. That will have the same cost irrespective of resolution, geometry density is what matters in this case. If most of that is baked and streamed from storage, the cost will be low. If it's dynamic, it could pose a problem.
**Ray tracing probes distributed in worldspace is resolution independent, but also is not an aproach used by any game so far.