Again you infer too much from my isolated remark that it be risky.
Please please don't draw conclusions too quickly. I know we extrapolate from any single bit of information to the whole picture. There must be a link here on LW about it. On my side I promise I will not post such isolated statements again.
To get back to your point. We prefer natural consequences too. My wife is better at it then I am. She immediatly has a consequence at hand that mostly does work. But then she has the plan for the day. A long time ago I agreed that we establish the same rules and back each other on decisions made by the other. But often enough I find myself in a situation where I have to back her - but don't know to what end. That places me into a situation where I am bound to fail.
As for the riskyness of the approach. I'd like to just point out the logical consequences: If you establish a rule (e.g. because if all follow the rule that will have positive effects for all) and then do not perform your part in supporting this common cause and instead look away then the same will happen as in any social group where social norms are not followed: the rule will weaken and possibly fall below a level where its usefullness gets negatve. If you want to maintain the rule you should better invest sufficient continuous energy to keep the rule adherence at a level where it is useful instead of letting it slip and then invest significant energy into getting it back.
But often enough I find myself in a situation where I have to back her - but don't know to what end. That places me into a situation where I am bound to fail.
To me that sounds like the rules that your wife sets a too complex. If you don't understand that they make sense as an adult, they will be completely arbitary to the children.
Followup to: Strategic ignorance and plausible deniability
My in-law always says: "For children it is easier be forgiven then to get permission."
EDIT: This post is superseeded by my Book Review: Kazdin's The Everyday Parenting Toolkit I recommend reading only that. The remaining insight of this post is: Children expend more brain power on their parents than the parents on them.
I can say from experience: That is risky.
Children (esp. small ones) expend significantly more brain power on their parents than the parents on their children (your mileage may vary). I can assure you that they will notice these cases - at least some - and take that into account one way or the other.
If the children notice this they may assume that you either condone, accept, bear or ignore it. None of these has positive effects.
Possible alternative strategies:
I am influenced by The Adlerian School. Of relevance here is Striving for significance.
The testing of limits and the resulting interaction with the parent give the child a feeling of significance if the parent acknoledges the act of the child even if he doesn't agree with it. On the other hand ignoring the act of the child is negative feedback about significance.
EDIT: The asymmetry between parents and children with respect to the effectiveness of deniability can be generalized to any situation where one actor has significantly less overall information about the situation than another actor and thus might not be able to reliably estimate whether deniability is possible.
ADDED: tadamsmar pointed out that ignoring is scientifically known to be effective and the advice or rather personal expierence I have related in this post may be contraproductive (at least if applied in isolation).