So in theory, low frequencies can be beamed almost symmetrically forward and behind the speaker (with reduced sideways radiation).
I'd still like to see real measured klippel data for an open baffle design to see how close they get to this theoretical ideal.
Edit: In the first few posts open baffle designs were championed for their "room filling" sound (which to me implies omnidirectional behavior, which would tend to more strongly excite room modes). It seems more accurate to state the possible advantages of the open baffle design as controlled directivity that reduces sideways radiation, particularly at low frequencies.
In open space - yes. In small rooms (all our rooms are small compared to low frequency wavelengths) it is much more complicated.
Once again I will doubt that Klippel is a proper way to measure OB at low frequencies because some space needed to sum front and rear sound, thus nearfield measurements can be confusing. But I posted DIY free-field measurements, I think they are ok to see whats going on.
I can agree with your last thesis. Good way to think about it in three different regions.
At higher frequencies, OB can add feeling of deeper space because of rear radiation and Haas effect, if you have enough distance to the wall and it can be a good thing. Or make unwanted reflections if the wall is close and it's bad. Side radiation can play a game too, but not very much.
In low-mids reducing of side radiation may be a good thing if you place speakers right, I illustrated it. Rear radiation not very problematic, but can create SBIR if speakers are close to the wall.
In bass (modal) region you can think of OB as more woofers in the room (one woofer acts like two with opposite polarity). This in many cases gives a more distributed modal picture and can produce more flat bass. But maybe not. It's just as dependent on placement in the room as it is with any other speaker.
And because OB often generates less bass compared to box speakers, with less modal problems, bass can sound very dry and fast. No bass — no problem