ORIGINAL: ChazM
Speaking of which, 1/4 point deduction increments has been tried here and seemed to help large swings in a maneuver's score and did show more agreeable scores by a judging pair.
Research in the field of testing theory suggests that most people are unable to consistently discriminate using a 10 point scale. That is to say, results using a 10 point scale turned out to be no more reliable that results using a 5 point scale. So in reality the difference between, say, an 8 and a 9 was really just illusionary.