Oftentimes new users innocently specify the winning criterion as fewest characters, as opposed to fewest bytes, without being aware of the differences. As a result some of the answers may be encoded into large Unicode values. But who enjoys seeing an answer solely consisting of a big Unicode turd? I propose that in the FAQ, site intro, or wherever, there should be a warning that it is better to use fewest bytes.
I would advise scoring-by-
I think scoring-by-
characters encourages weird base256/unicode compression of the source, which IMO generally doesn't really add anything to the contest, and merely serves to obfuscate the code, making less accessible answers, thus decreasing the overall quality of the site.