There are some issues in the UI that bug the heck out of me:
I'm running 1.1.0 (final).
+1 on these issues, especially the non-persistent 'passed' and 'skipped' options.
Regarding my second point; it looks like what is happening is that not only failing test are still executed, but just all tests are always executed, regardless of selection or not.
the "other" specs should be skipped, not actually run. But since Jasmine doesn't stop a spec on failure, it is possible to have more failures than there are specs. Which might be why you're seeing odd counts.
The behavior around skipped and pending specs on master - and what's targeted for 2.0 is much more obvious behavior than before.
If you see this type of bevhavior on master, please open a new issue. Thanks for the feedback. This issue helped a lot with spec execution behavior for 2.0.