When the rules of a role-playing game are broken, one general purpose way of fixing the rules is to rely on the players to be reasonable and not to take advantage of the broken rules. I thought I'd give my opinion on the usefulness of this technique.
In general, this technique relies on having players who are not power gamers. A player who is a power gamer will naturally want to do whatever is most powerful within the rules, and will not normally want to hold back in any way, so this technique will not work (unless they agree to totally change their play style). You can try to make it work by threatening to punish the player for abusing the rules, but I find this sort of player-GM interaction to be highly undesirable. Power gamers are likely to be happier if you either fix the rule or let them exploit it. I will assume for the following cases that none of your players are power gamers.
Relying on the players not to abuse the rules works best when you have to perform unusual actions to take advantage of the broken rule. It works even better if the decision to abuse the rule can only be made during character generation. For instance, a classic way to break the rules is to find a combination of multiple powers, which, when taken together, become a devastating combination. If the players are not trying to power game and are just making cool characters, chances are that no one will ever choose this combination, and once play begins, this combination will not be available for use. Therefore it won't, in practice, create any problem at all for your game. If everyone in the group knows the combination is broken, they may even decide to intentionally avoid it for that reason.
Another case where it is very typical and useful to rely on the reasonableness of the players is when you have multiple choices with different game effects and different special effects, and you want the choices to be balanced. Ideally, each choice would be exactly game balanced with every other choice. In practice you cannot always achieve this, but if the choices are reasonably close to being balanced, it will not be a problem. Players who play "in character" will naturally want to do things that are fun and appropriate. If option A is more fun and appropriate than option B, and is not blatantly inferior, they will choose option A even if a deep analysis reveals that option B is probably a bit more effective overall. After all, the purpose of a role-playing game is to have fun, not simply to win at all costs. When designing game rules, I often rely on this. If a rule is close to being balanced, I can just assume the players won't try to power game and move on to the next rule, rather than agonizing over every little detail.
On the other hand, if option B is clearly far better than option A, you have a definite problem. There is no getting around the fact that the game balance is flawed. It is certainly possible for players to stay in character and choose option A anyway, but I think this is not as good a solution as fixing the rules. First of all, it puts a lot of pressure on the players. It can be awkward to make a decision knowing that it is making your character far weaker .Even if the players are not trying to abuse the rules, if the rules are not balanced, you end up with a situation where some characters and actions are far more effective than others for reasons that don't make sense within the story. For instance, it is very annoying when you want to make a character who is a mighty warrior, but because your character conception leads you to choose powers and abilities that are inefficient in game terms, you end up being a rather mediocre warrior who just tries to provide support for his more powerful buddies. Also, rules that are broken in terms of game balance are often (though not always) less fun to use. A power that is broken on the “too strong” side can seem bizarre and ridiculous; a power that is broken on the “too weak” side can be boring because it has no effect.
Another problem is when you have a slippery slope, which is the case with the daily powers problem I mentioned in the last post. If it is efficient for the party to fight fewer encounters per day, and there is no definition of how many you "ought” to be fighting and no real penalty for resting too often, then basically the players are being told to just decide for themselves how powerful and successful their characters will be. They can rest more or less often, and are always tempted to rest more often except for the feeling that it doesn't seem right. Giving the players total control can work under the right circumstances, but this seems more like interactive storytelling, where you don't really need rules at all. If you are going to use a game with rules, I usually I think it is more interesting if the game sytsem actually provides some sort of meaningful challenge to the players, some sort of framework that the players can interact with so that they are playing a game in addition to telling a story.
Of course the players (or the GM) can come up with self-imposed limitations – “we must fight 4 encounters per day”. Having guidelines can be very helpful in terms of reducing uncertainty about what is abusive. But to the extent that the guidelines are followed, this is really an informal way to make new game rules rather than a way to avoid having to make new rules at all. And to the extend that the guidelines are not followed, you still have the uncertainty of when and why you shouldn’t follow them.