I voted for the lowest scores … not because I’m not serious (well, in fact I am not, lately I am only playing with it from time to time) but because I find more practical to maintain more smallest projects, per-feature, per-story, per-scope etc, than a single large one.
Plus, I have my own version of a “suite” – I call it a “Suite Test Case” (STC) managed by groovy code and supported by my own settings stored as JSON in the properties/Description field in the individual TCs.
If an STC detects before-tc or after-tc properties, then chaining of the relevant TCs occurs too.
We have 2 Production environments and 4 test environments but test cases are designed to work on all environments using profiles. Test Suites are in form of feature stories and then there is one big test suite that contains each and every test case 300+ that runs whenever a build is delivered.