You should use whatever is cheapest and best known for your platform and call it a day. Obfuscation of high-level languages is a hard problem, because VM opcode streams don't suffer from the two biggest problems native opcode streams do: function/method identification and register aliasing.
What you should know about bytecode reversing is that it is already standard practice for security testers to review straight X86 code and find vulnerabilities in it. In raw X86, you cannot necessarily even find valid functions, let alone track a local variable throughout a function call. In almost no circumstances do native code reversers have access to function and variable names --- unless they're reviewing Microsoft code, for which MSFT helpfully provides that information to the public.
"Dotfuscation" works principally by scrambling function and variable names. It's probably better to do this than publish code with debug-level information, where the Reflector is literally giving up your source code. But anything you do beyond this is likely to get into diminishing returns.
