Both types have their adherents, obviously, and there've been some interesting threads on the topic before. Search the archives and you'll find them.
When I chose my TT I was also unable to audition. I chose by following this (admittedly deductive) "logic":
- making a good suspension is both difficult and expensive (I had an inexpensive, suspended deck as a negative example)
- assume a price point (let's say < $10K)
- consider a table of each type at that price point
- the maker of the suspended table must dedicate a significant portion of his costs to designing, engineering, sourcing and building the suspension
- the maker of the non-suspended table dedicates virtually none of his costs to the suspension (just three cones or whatever)
- therefore, the maker of the non-suspended table has more resources available to optimize other critical elements like the bearing, motor, plinth and platter design and materials, etc.
Therefore, this hypothetical non-suspended table "should" outperform its similarly priced hypothetical rival, and will certainly be easier to set up and forget.
It seems likely that a non-suspended table should be more dynamic. The less the table moves laterally in response to energy from groove transients, the more of that energy will be transmitted to the cartridge generators. OTOH, a good suspension should isolate the stylus/groove interface from structure-borne and motor vibrations better. This would yield blacker backgrounds and a lower level of detail retrieval. Nonsuspended tables rely on higher mass and a mix of materials to meet this goal.
This is all I had to go on when I made my decision. I'm happy with the results on my Teres, but so are most Basis owners!