I don't like the term 'testing of limits'. It seems to be imply that somehow children are hell-bent on testing the boundaries of the acceptable and doing the naughtiest things they can do without getting caught, just for the sake of being naughty.
In my experience children do things like that. I have two little brothers and they had phases where they did engage in a lot of testing of limits.
Testing the limits is neccessary to learn the real rules. Especially in the beginning they do not even know what a rule is and even less what consequences to rule violation are. These are higher order concepts which must be learned.
Same with defiance and anger: To make any use of emotions their effect must be calibrated. The brain surely is not hard-wired to associate speific social situation with specific emotions. Not when the social situations themselves are largely learned. I think that a lot of anger and fear etc. is triggered during these childish ou...
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).