To continue the conversation we were having on the
Post production panning thread I agree that the technically true ambisonic reverbs can be underwhelming. As with all things in sound design there is the question of reality verses expectation. A good stereo reverb has a long pedigree of meeting the expectation of how a room sounds even though it is usually a long ways from the reality of how a room sounds. I think the desired illusion of a really immersive reverberant signal over 2 or 5.1 speakers leads to some wild and creative techniques that you are leveraging with your pre-set. Ultimately this is good because ambisonics has come to maturity in an almost exclusively academic setting where the reality is almost always the goal.
I used the same technique for decode that you are using when I created an ambisonic web-player in the early days of the web audio api. I was told that although it's not completely accurate from a mathematical perspective it was inspiring to see it done. As non-academics I think inspiring (and entertaining) is what we aim to do.
Incidentally I have also had good results with a single stereo reverb run through the ATK superstereo and then pushed/focused into the section of the sound field where I want it. This way I can paint different reverberant sounds to different regions of the sphere.