The SP10MkIII can keep the Timeline's dot on the wall all day long with repeated playing of the LP uncaring of multiple needle drops. Control theory being what it is, the precepts were well understood in the 1980s when the MkIII was designed and Technics had all the expertise needed to do the job right and they did. It is certainly true that computers can now be used as effectively since that is the big thing that has changed in the last 35 years and that is what Technics is now doing in their newer machines.
But their specs are no better, and FWIW this is because the original servo was properly designed and executed, which is also why the MkIII made an excellent platform for LP mastering.
Unless you can show me that the servo in the older machine is actually hunting or some such, I'm not taking it on anecdote that what you are saying about the older Technics is real. The specs Technics publishes don't seem to agree with you as far as I can make out (but giving you some benefit of the doubt, its also obvious that the actual machines outperform the published specs by a fair degree, so its an unknown as to whether Technics has made any improvement with their new machine).
Now I get that you had a machine but if you are saying you heard something you like better I've no problem with that, but please keep in mind that is a nice anecdote. One variable you did not address is that the stock Technics platter pad, which is supposed to control resonance in the LP itself, does not do that very well. If you did not use the same platter pad on both machines used in a comparison, the results of the comparison are quite suspect!
The one concern I would have in all this (I don't really have a dog in the fight) is that the older machine is subject to the whims of age: electrolytic capacitors can fail, resistor values can drift and corrosion can attach rare semiconductors, compromising them or causing outright failure. But if the refurbished machine is kept in a controlled environment and if replacement semiconductors are around it should be fine.