A bad DM would be bad regardless of if the rules supported them taking away a PC's powers or not.
This ignores the possibility that rules can
help, by saying that unless rules are perfect defenses, they are useless. This is a perfect solution fallacy: since the solution isn't perfect, it is necessarily useless and should be ignored.
Rules can do things like:
- Make it much more obvious when a GM is behaving nefariously
- Give players a justified standpoint for providing criticism
- Help well-meaning but inexperienced, unwise, or wrongly-taught GMs see the error of their ways
- Help a "bad DM" realized that they are being a "bad DM" and, thus, potentially desire to change
All of those are worthwhile things to have happen, even if rules fail to provide an absolute, perfect defense against GM misbehavior.
Now, of course, I am speaking in hypotheticals--things rules
could theoretically do. So it is valid to say, "Well, not
all rules can do that, so this doesn't prove rules are
better." And that would be correct--but I'm not arguing rules are guaranteed always better. What I am arguing is that this attitude--that because rules cannot perfectly shield against bad GM behavior, they're worthless for doing anything about bad GM behavior--is wrong and harmful. There
are tools we can employ which would help. There
are structures that could make significant progress, and even (ideally) actually help convert a few bad GMs. It is worth our time as players, and as (armchair) designers, to talk about how rules can
help, even if they'll never
solve.
After all, isn't that exactly the argument folks like
@Micah Sweet have made in defense of pursuing verisimilitude? There is no ruleset that can achieve perfect verisimilitude--it is provably impossible, even--but that is no excuse for abandoning the quest for it. It just means we must temper our expectations, and focus our efforts on the places which achieve the most gain for low and/or acceptable cost.