Suggesting skating forces cannot change the cantilever angle vs. saying the changes will be minimal and not dominate an individual situation are two very different things. One is a seemingly incorrect statement of fact and the other an opinion and it is unclear which you are trying to propose.
The conflict that is happening here is the validity of a single parameter (TAE) being used as the only metric that is important when comparing various arm geometries. You have already addressed how other factors make the complete elimination of TAE of a LT arm a 'bad move' . You then revert to the single parameter to discount the large population of people here who have found that underhung arms sound good in spite of the drastically higher TAE. This completely ignores all of the other possible explanations. Playing both sides of the fence when it suits you is not accepted scientific practice.
I set out to prove to myself that TAE was the dominant parameter and find that ~3° of TAE at the inner groove was unlistenable in a traditional offset arm and somehow far greater values with an offset free underhung arm was musical as hell. It was this disconnect in addition to the numerous other reports of UH arms sounding way better than the gross TAE would suggest that started me wondering if there was something else at play.
I asked above what underhung arms you have experimented with and would also be curious as to where the single null point was placed in your listening tests. If you have experimented and found that the sonic results of similar levels of TAE error is consistent across all of the possible arm geometries then you are free to voice your opinion.
dave