Hi Steve -

Thanks for posting.

This is a really, really common problem. There are a couple of things that can often cause it:

1. The tests are sequence dependent. This means that the execution order can affect their outcome. For example, when one test is run, it may change the state of some resource (maybe a static member or database) in such a way that prevents other tests from operating correctly if they are executed in the same process. See (test atomicity for more information). Try running the tests in a variable sequence when re-using the task runner process (options under the 'Advanced' context menu).

2. The tests may have intermittent pass/fail tendencies. This is often caused by multi-threaded behaviour or reliance on things outside the test runner process that may have an inconsistent state (see multi threaded tests for more information).