Thus the strategy that can "choose" a more effective form won't have to engineer a new version on the spot if there's an appropriate one already in the bank. It won't require that long, long wait for a chance mutation either. Which doesn't preclude such events from happening - but nevertheless these strategies, if any good at all, will have provided other options for themselves as a consequence of their own evolvement.
(I'm reminded that Hermit Crabs choose empty shells as protective structures - and the shells weren't there for the expectant purpose of being chosen by a crab. But of course this could be consistent with either of the rules in play here. Except we see that strategies are essentially opportunists.)
So if it looks to neo-Darwinists that the accidentally acquired longer neck of the Giraffe gave an advantage to an existing strategy for eating leaves, and that strategy adapted itself to the new neck, that might ( theoretically) have happened. Except the mistake has been to rule out other roads the adaptive process could have taken. Such as the one I recommend above.
And we are already learning that there are preadaptive elements in our genomes that may be there for just that purpose - to avoid the wait for randomness to make a structural alteration that can (and not just theoretically) be arranged in advance in anticipation of its future usefulness.
(Certainly these variations were in wolves that dog breeders used to form new varieties of dogs, fitting the breeder's strategic goals to wolf/dog primary strategies in the process. Manufacturing "learned" behaviors that, voila, became almost immediately heritable! And don't tell me the wolf was always man's best friend somewhere within its backup planning).
Aphorism of the day: Experience drives the evolution of strategies which apply new "knowledge" to the evolution of their forms.