My understanding is that the listening room, at least in its current form, allows you to adjust distance to the sidewalls and that it supports rapid switching of stereo and three channel setups as well:
My understanding is that Sean Olive's study, published in 2004, used the 21.6 foot wide room exclusively. Nor am I aware of any subsequent single-speaker studies where they used a narrower room. In that 2004 paper, Sean lists four limitations of the study. The first one is this:
"Up to this point, the model has been tested in one listening room."
There's also a 1985 Toole paper that was conducted with the speakers placed at the corners, and this was a study of directivity specifically:
View attachment 81782
So I don't really think the matter of the Toole/Harman apparent inclination towards wide-ish directivity has much to do with the listening setups.
I agree that in Toole's 1985 paper the sidewalls are close enough that early sidewall interactions are included. I was talking about the tests done by Harman in their 21.6 foot wide speaker-shuffler room, hence my wording.
I think the caption for the diagram of Toole's 1985 test includes incorrect information: To the best of my knowledge the Quad 63 WAS NOT "equipped with absorbing pads on the rear half of the enclosure to attenuate the output above 500 Hz." It was the original Quad, the "57", which was. I have owned both. Otherwise the description matches the ESL 63.
In my opinion the Quads were handicapped by having the high frequencies in their backwave energy absorbed by the drapes. One of the things the 63's get right is that the backwave energy's reflections are spectrally correct, and the drapes prevent that.
The Definitive Tech speakers with the extra midrange and tweeter facing the rear have a separate level control for the rear-facing drivers. I suspect that for this to live up to its potential you need a very large room such that the speaker are a good distance from the wall behind them. In most homes it probably does not work at all well, which is probably why those speakers weren't loved by a lot of people. I suspect that most people who bought them gradually turned down the rear drivers over a period of time until they were eventually silent.
I have manufactured controlled-pattern bipolar speakers in several variations, and as you suspect the distance from the wall matters... or more precisely, the amount of time delay before the arrival of the "backwave" matters. I recommend enough distance to get at least 10 milliseconds time gap between the direct sound and the arrival of the backwave energy, which normally translates into about 5 feet from the wall. My more recent models aim the rear-firing drivers upwards somewhat which decreases the distance from the wall needed because the reflection path now also includes a ceiling bounce.
We have conducted in-house blind listening tests and find that there is a "sweet spot" regarding how loud the backwave energy should be. If it's too loud, clarity starts to be degraded. But at levels just below the clarity-degradation threshold, envelopment and timbre are enhanced with no audible detriment. IF the clarity-degradation threshold is exceeded, then there are tradeoffs involved. In my opinion rear-firing drivers have the potential to improve both the spatial and timbral qualities, when done right... but "done right" implies designing with their contribution in mind from the outset, rather than tacking them on as an afterthought.