Yes Simon, you have raised valid points. I will add more explanation, hope my responses are only improving the status of the discussion (than hindering it).
1). So when someone is refactoring, if a testcase is removed; then author is loosing only the mark that was previously granted for the testcase addition; so this might be reasonable. However, as you correctly suggested if the one who added the testcase and deleted the testcase are two users, then it is kind of a problem since +1 was given for someone, but reduced from someone else.
If we are not going to remove that already granted mark, the symmetry is gone. How should we correct this? Any suggestions...
2). Here also, when a test was added we granted +1, then when test started failing we granted -2 against that user. So I thought of granting just +1 mark when this failing test is removed, so that the author is finally receiving 0 marks for the whole trip. (However if he/she fix a failing test, +2 will be granted).
It would be great if we could decide on what points to be granted on each of the 12 scenarios (+ any missing), so that we can resolve the asymmetric marks granting issue.