In my test list, my ordered test appears and runs, but each individual test that composes the ordered list is run a second time outside the list if I run all the tests.
What is the point of the ordered list if it's going to run them nondeterministically as well? Am I missing something?
For unit testing, ordered tests is exactly what you DON'T want.
However, for some cases of integration testing and regression acceptance testing where you do automated end-to-end tests of a system you might want it. For instance, application test data often is expensive to prepare and you might decide that some limited side effects between tests are ok, but should be keept at least consistent between test runs.