Nice. Did they let you see their equipment? I'd bet my cinema was cal-ed to near 82 dBC or maybe a tad lower as well. It was noticeably lower than I play stuff in my home these days, which is around 80 dBC, albeit with the influence of the much smaller room. With the thinness and slight 2 kHz push at the cinema though, I'm glad it wasn't played louder. Did they give you any info on how they did the cal? Was it X-curve all the way? Or did they let the low freqs rise a bit?
Assuming the cinema I went to did a by-the-book X-curve cal, I'd guess that the dub-stage was on the dry side and didn't appreciably build-up room energy until below 200 Hz or so. Or else, they didn't do much EQ above there for other reasons.
I think a lot of newer dub-stage / mix spaces are getting thicker absorption than in the past. A striking example of this is the "Game of Thrones" TV series. I believe their studios follow ATSC recommendations and calibrate magnitude-smoothed or power response to a flatish curve, which suppresses the bass compared to an anechoic flat system optimized for music. To correct for the bass boost that they are likely adding, I use with a low frequency shelf somewhere. For seasons 1-4, my shelf was centered around 265 Hz or so, IIRC. Results were a bit inconsistent from episode to episode suggesting that they were either working between multiple rooms or were addressing the low end deficiency in different ways each time. Then for season 5, they started doing Atmos mixes, which were almost certainly done in a new purpose-built studio with more low frequency absorption. The 265 Hz shelf sounded all wrong, and I had to move the shelf down to 225 Hz or so to get all the dialog to sound balanced again.
I'm sorry to hear about your terrible LieMAX experience. I'm going to take a wild guess as to what may have gone wrong. AFAICT, Audyssey tech fits magnitude-smoothed response to a curve rather than power response, which is used for the X-curve. I've noticed that magnitude-smoothing, even at high resolution, actually omi a lot of tail-end reverb energy in room measurements. I believe the error is actually greater for higher frequencies because of the frequency-varying time-localizing effect of the smoothing kernel. E.g., 1/48th octave at 10 kHz is a lot more time local than 1/48th octave at 500 Hz. So as a consequence, the high frequencies may appear lower in a magnitude-smoothed measurement than a power response measurement, and XT32 may be boosting them too much. Even just a dB or two hot up top is a recipe for some nasty overly bright sound. You'd think that the people agreeing to license the tech would bother to listen to the result and decide that there's something wrong if it's painfully loud at "normal" playback level. But that's not how technology works these days.