I've had success with Vishay CH02016 resistors up to 30 GHz on 7.9 mil thick Megtron 7. Potentially you could make it work to 40 GHz with careful matching and accounting for solder thickness in your 3D model.

Dot terminations work well too, but I've done those on ceramics (Alumina) and they are high-pass.

I'd take a look at SAW filters. Try: https://www.taisaw.com/

I've been documenting a lot with Jupyter notebook because it uses simple markdown (so it does simple formatting and images). It also executes code -- for example, I find myself plotting s-parameters using scikit-rf a lot. It can be exported to a single HTML document for sharing with others.

Seems like a lumped element solution might be possible at 1.5 GHz. Maybe "<5 dB insertion loss" means that the 5 dB in addition to 6 dB loss from the 4-way split.

Yes, I think that is correct. Your transformer will be made up of transmission lines with impedances of 50.459, ..., 297.217 ohms. I believe an impedance of 297 ohms might be hard to realize in microstrip or stripline. Maybe it has to be twin line?