hello. i'm sorry to say it, but the neat answer isn't as new as you would expect... it has been published in at least five books (however, it isn't very popular in the Internet, if it makes you feel proud ). just take a look at http://www.cut-the-knot.com/blue/OddballProblem1.shtml. however, i'm not accusing anybody of plagiarism--this solution had been found many times independently (quite a good exapmle of convegence, you would say).
the problem with this solution is the rules of inference cannot be observed so clearly as in the ``classical'' solution. it would be quite good if someone (the author?) explained what is the rule of putting the balls on the left and on the right throughout the 3 weightings.