Oh, I think I understand quite well that when Bazant says Pdyn/P0 ~ 31, he is simply saying that the dynamic force was 31 times greater than the design load. That does not imply or "theorize" that there should have been a 31g impulse and 31g deceleration in the upper block, unless the columns below were actually capable of resisting 31 times their design load! If anything, Bazant is demonstrating the absurdity of your interpretation!
Indeed, Bazant's very next sentence sentence makes that clear:
In spite of the approximate nature of this analysis, it is obvious that the elastically calculated forces in columns caused by the vertical impact of the upper part must have exceeded the load capacity of the lower part by at least an order of magnitude.
The load capacity of the upper part was no greater than the load capacity of the lower part, so Newton's third law implies the load capacity of the upper part was also exceeded. That's obvious enough to be left unstated, but Tony thinks Bazant and Zhou were unaware of it.
You apparently also can't understand that when Dr. Bazant says
P = Pdyn yields the following elastically calculated overload ratio due to impact of the upper part: where Po = mg
and he calculates Pdyn/Po as 31, that this means Pdyn is an amplification of the static load (mg) of 31 times meaning a 31g deceleration would have had to occur for that to happen. I can understand your confusion somewhat as Dr. Bazant also calls Po the design load capacity but it isn't. It is just the static load. The design load capacity would be the static load x factor of safety. The 31g amplification would thus be an overload of 31/factor of safety.
That part of Tony's argument makes no sense. Bazant and Zhou's estimated force ratio of 31 (a dimensionless number) does not imply a deceleration of 31g (with the units of acceleration) would have been achieved, and it certainly does not imply a deceleration of 31g would have been achieved for the length of time necessary to create a jolt of the magnitude Tony has imagined. To calculate the deceleration, Tony will have to estimate the duration (or a related distance from which the duration can be calculated) through which the decelerating force is applied.
Tony goes about that backwards. In the cited early version of Tony's paper, he estimates that the velocity should be reduced by 13 ft/sec; that's the size of his jolt, and it's basically the first thing he calculates. From the size of the jolt he expects, and from his baseless assumption of 31g, he calculates a 13 millisecond duration for the 31g deceleration. Even if that weren't going about it backwards, neither the upper section nor the top of the lower section is strong enough to exert anywhere near the amount of force needed to achieve a deceleration of 31g. (As I calculated earlier in this thread, using Tony's questionable numbers, those sections aren't anywhere near strong enough to achieve a deceleration of 10g, let alone 31g.) So the structures will fail long before they can achieve any more than a small fraction of that deceleration, which means the true deceleration will be nowhere near as large as Tony believes.
According to my calculations, using Tony's (questionable) numbers, the portion of the jolt attributable to the load capacity of the lower section is
insignificant compared to the portion that would be attributable to conservation of momentum in an inelastic collision of two perfectly rigid objects.
The fraction of velocity lost to conservation of momentum is easy to calculate for Tony's idealized model, because it's essentially the ratio of the mass of the topmost floor of the lower section to the total mass of the upper section: less than 10% (and less than 2 ft/sec). For reasons that have already been discussed, it is doubtful whether Tony, who has a hard enough time seeing an obvious tilt, would be able to see a less-than-10% reduction in the velocity of the upper section's lower "edge" by looking at the roof line in the videos. He would have a hard time seeing that small a jolt even if the upper section were perfectly rigid and were to strike the lower section perfectly square---and those are not realistic assumptions.