Part of the problem with CERA is that the quality of the pitchers caught plays a big part in the stat. If someone caught Weaver 5 times and was a crappy catcher, and caught Blanton 5 times and was the best in the game, CERA will show the crappy catcher as the better one.
While the theory behind CERA is a good one, it needs to evolve somehow. Taking into account what the pitchers ERA is, and what his ERA is when Catcher X caught him. Maybe a CERA differential? Going into the game, pitcher had a 4.00 ERA. In 6 innings he gave up 2 runs. So for those 6 innings, he CERA differential is + 0.668, calculate for the relievers and so on. Then again, maybe not, since if a catcher caught every inning of every game, he should get a differential of 0.0. /shrug.