As long as the resulting vector is in the wanted direction I don’t think there’s much point in mirroring the setup in any way for the purpose of achieving “balance” as any such should be able to be defined by the shape of the reflective surface and the fuel expansion properties (is this just counter-intuitive or am I wrong? Perhaps think of it as if it was an aerospike engine for use along a hypothetical sharp border between atmosphere and vacuum: there would only be a point in having a properly “tuned” custom “half annular aerospike” that dipped into the atmosphere just like a boat engine).

“Ideas are cheap, implementation is hard.” Henry Vanderbuilt

The opacity of hydrogen does increase quite a bit once the hydrogen is ionized, but 1) that means making it pretty damn hot, 2) the opacity is still small (less than 1 cm2 per gram of hydrogen), and 3) the opacity is principally *scattering* rather than the desired absorption. For heating hydrogen directly, I imagine your better off with an RF beam, and you’d probably want to introduce some some magnetic fields and start using plasma effects. It rapidly distances itself from the simple concept you propose.

If it is possible to use this concept as intended, then it should be better than either. Please note that I said if. Higher Isp and T/W than nuclear, and much higher T/W than electric anything at probably 40% of that Isp. For acceleration and fast transit, a clear (note IF) winner on transit times.

Nuclear thermal Isp is limited by the temperature limits of the heat exchanger. Since the hydrogen must be heated by a physical heat exchanger, there are serious limits to the temperatures attainable. If the hydrogen with a molecular mass of 2 could be heated to the same temperature as a stochiometric chemical H2/O2 engine exhaust product with a molecular mass of 18, then it could have an exhaust velocity 3 times as high.* That would be Isp=~1,350. Actually less because the chemical engines run rich which lowers the average molecular mass.

Anything electric has powerplant mass issues that kill T/W to the point of uselessness for reasonable acceleration and fast trips. IMO of course.

The nuclear material temperature limits are what got me thinking about this afterheating trick. A laser from dead astern would be an ellegant solution, unless you have to carry it and all it’s gear along.

*I’m sure you are aware of the exhaust velocity being proportionate to the square root of temperature over molecular mass. Just letting you know that’s where I got my numbers.

How fast is an unanswerable question because the question is incomplete. The simplified equations that apply are
F = m a
v = u + a t

To get v we would need to know the mass m of the rocket, the force F (also called the thrust) and the burn time t. The mass changing as fuel is burnt makes everything harder. Given sufficient thrust and fuel we can make the rocket go as fast as we want (until relativity kicks in).

How fast? Propusion concepts like nuclear thermal and VASIMR promise dramatically-reduced transit times to Mars and beyond. For manned missions that is a big plus.

Am I asking the wrong question in the wrong place, perhaps? The concept in this blog post of yours may work for slow transit times for unmanned payloads.