Looks pretty simple to me. You cannot say, when you have evidence of the existence of dogs 33,000 years ago, that dogs were first domesticated 11,000 years ago, so starch-related changes post-date domestication by 22,000 years.
Equally, once you have established that starch-related changes post-date domestication and for that matter so does the human adoption of agriculture, you need another excuse for domestication.
Skulking round rubbish tips, a practice adopted by not only wolves but jackals, foxes, rats, mice, gulls, crows, etc, etc, is NOT domestication and given the list I've just recounted, does not lead to it, either.
While the route to domestication could be either commensal co-operation or pupnapping and training (or both) and we may never know which, the reasons for domestication must be advantage to humans and to find those you look at the areas where wolves (dogs) score over humans: hearing, smelling, running, biting.
This gives you a range of jobs for the new addition to the human armoury: guard dog,tracker dog, hunting dog, sled puller( last resort meal). Only one or just perhaps two of these are an advantage to a sedentary agricultural population whereas hunter-gatherers or nomadic pastoralists can utilise all of them.
Of course nomads are less likely to leave identifiable fossils in places where they are likely to be found, so the early dogs - which in any case are likely to be difficult to separate from wolves on e.g. skull morphology (if you don't believe me, read some mid-twentieth century literature on wolf sub-specific classification) - are less likely to register in the areas where they are more likely to have originated. So finding earliest dogs in areas with sedentary human populations becomes a self-fulfilling prophecy.
John